From 6070f9b778b0754b229dddd7a9bd1711502deede Mon Sep 17 00:00:00 2001
From: "Daxiong (Lin)" <192523189+comfyui-wiki@users.noreply.github.com>
Date: Mon, 16 Nov 2026 23:24:57 +0800
Subject: [PATCH] docs(pricing): sync partner node rates
---
es/tutorials/partner-nodes/pricing.mdx | 490 ++++++++++++++++--------
ja/tutorials/partner-nodes/pricing.mdx | 491 +++++++++++++++++--------
ko/tutorials/partner-nodes/pricing.mdx | 491 +++++++++++++++++--------
tutorials/partner-nodes/pricing.mdx | 449 +++++++++++++++-------
zh/tutorials/partner-nodes/pricing.mdx | 491 +++++++++++++++++--------
5 files changed, 1680 insertions(+), 732 deletions(-)
diff --git a/es/tutorials/partner-nodes/pricing.mdx b/es/tutorials/partner-nodes/pricing.mdx
index fa6304f036..3d3e3442b1 100644
--- a/es/tutorials/partner-nodes/pricing.mdx
+++ b/es/tutorials/partner-nodes/pricing.mdx
@@ -3,54 +3,56 @@ title: "Precios de los Nodos asociados"
description: "Precios en créditos de los Nodos asociados de cada proveedor (Anthropic, OpenAI, Kling, Luma), por modelo, tipo de token y tipo de generación."
sidebarTitle: "Precios"
mode: "wide"
-translationSourceHash: 2ef8ce31
+translationSourceHash: f68d51fd
translationFrom: tutorials/partner-nodes/pricing.mdx
translationBlockHashes:
"_intro": 1ec20604
- "Anthropic": 61f55fec
+ "Anthropic": 1757fac0
"Beeble": d1b19f61
- "BFL": 7ae11dc6
+ "BFL": 0a63f1b4
"Bria": 94080005
- "ByteDance": d2af4769
+ "Breeze": f38692c6
+ "ByteDance": fec72d04
"Clarity AI": 3d568b52
"ElevenLabs": cadb48cc
- "FishAudio": c957b7da
+ "FishAudio": 24ec3ea0
"Magnific": 4000e7e1
- "Google": b8be5b63
+ "Google": 958913c2
"HappyHorse": bceb584a
- "HeyGen": 7f08ab4e
+ "HeyGen": 248354a7
"Higgsfield": eed76dcb
"Hitpaw": 320c0537
- "Ideogram": da793624
+ "Ideogram": de0981df
"Krea": bcad7ba2
- "Kling": 32a38b1d
- "Lightricks": 3ba8d019
- "Luma": 9f809a47
+ "Kling": db91f830
+ "Lightricks": bbcb475e
+ "Luma": a2f2a8de
"Meshy": c057a289
"Meta": a253fe9a
"Mirelo": 2055c18a
- "Minimax": 07381eba
- "OpenAI": 17e24f34
- "OpenRouter": ef37d425
+ "Minimax": 3f9b84c5
+ "OpenAI": e51a94f8
+ "OpenRouter": 3de75ece
"Patina": d489c37e
- "Pixverse": bdb9b1c9
- "Quiver": 5538ce9c
+ "Pika": 66ba8620
+ "Pixverse": 86f21b3f
+ "Quiver": f84ffa99
"Qwen": 22dbe478
"Recraft": fc591c24
"Reve": 02eb39a9
"Rodin 3D": 2eb41834
"Runway": bd14d7ff
- "Sonilo": 48b9c581
+ "Sonilo": 66cc2806
"sync.so": 590402a9
"Tencent": cbe22aa6
- "Topaz": 5d08cf76
+ "Topaz": f557fb17
"Tripo": 842e861e
- "Pruna": f1b6a267
+ "Pruna": 08c63f88
"VEED": 1818bad3
"Vidu": be19d9ca
"Wan": 99df0ce3
"WaveSpeed": 349a7c9d
- "xAI": 28ff86de
+ "xAI": 6efd8650
"Cloud GPU": e228bb68
---
@@ -79,6 +81,7 @@ Los Nodos asociados se facturan por llamada en créditos. El tiempo de GPU de Co
| :--------- | -----------------: | --------------------------: | --------------------------: | ---------------------: | :------------------ |
| Opus 5.5 | 1.207 | 1.509 | 2.414 | 0.121 | 6.035 |
| Opus 5, Opus 4.8 | 1.509 | 1.886 | 3.017 | 0.151 | 7.543 |
+| Fable 5.2 | 2.414 | 3.017 | 4.828 | 0.241 | 12.069 |
| Fable 5.1, Fable 5 | 3.017 | 3.772 | 6.035 | 0.302 | 15.087 |
| Sonnet 5.5, Sonnet 5 | 0.603 | 0.754 | 1.207 | 0.060 | 3.017 |
@@ -104,17 +107,19 @@ Las imágenes opcionales agregan tokens que se facturan con las mismas tarifas d
### Imagen
-| Nodo | Créditos |
-| :------------------------- | :---------------------------: |
-| Flux 1.1 [pro] Ultra Image | 12.66 / run |
-| Flux.1 Kontext [pro] Image | 8.44 / run |
-| Flux.1 Kontext [max] Image | 16.88 / run |
-| Flux.1 Expand Image | 10.55 / run |
-| Flux.1 Fill Image | 10.55 / run |
-| Flux Erase Image | 6.33 / run + 0.84 / extra MP |
-| Flux Virtual Try-On | 7.91 / run + 1.06 / extra MP |
-| FLUX.2 [pro] | 6.33 / run + 3.17 / extra MP |
-| FLUX.2 [max] | 14.77 / run + 6.33 / extra MP |
+| Nodo | Créditos | Créditos con imágenes de referencia |
+| :------------------------- | :---------------------------: | :-----------------------------------: |
+| Flux 1.1 [pro] Ultra Image | 12.66 / run | — |
+| Flux.1 Kontext [pro] Image | 8.44 / run | — |
+| Flux.1 Kontext [max] Image | 16.88 / run | — |
+| Flux.1 Expand Image | 10.55 / run | — |
+| Flux.1 Fill Image | 10.55 / run | — |
+| Flux Erase Image | 6.33 / run + 0.84 / extra MP | — |
+| Flux Virtual Try-On | 7.91 / run + 1.06 / extra MP | — |
+| FLUX.2 [pro] | 6.33 / run + 3.17 / extra MP | 9.50 – 31.65 / run + 3.17 / extra MP |
+| FLUX.2 [max] | 14.77 / run + 6.33 / extra MP | 21.10 – 65.41 / run + 6.33 / extra MP |
+
+Al conectar imágenes de referencia se aplica la tercera columna. Solo con texto, **FLUX.2 [pro]** cuesta **6.33 créditos** más **3.17 créditos** por megapíxel de salida extra (`ceil(W × H / 1 MP) − 1`). **FLUX.2 [max]** cuesta **14.77 créditos** más **6.33 créditos** por megapíxel de salida extra. Con imágenes de referencia se añade **3.17 créditos** en [pro] o **6.33 créditos** en [max] por megapíxel de entrada, redondeado al alza: se suma `ceil(W × H / 1 MP)` de hasta 8 referencias. La tercera columna es el intervalo de salida de 1 MP para 1–8 megapíxeles de entrada de referencia (**9.50–31.65** / **21.10–65.41**), más el mismo cargo por megapíxel de salida extra que la columna de solo texto. La unidad **3.17** de [pro] está redondeada; el intervalo publicado **9.50–31.65** es el extremo de la insignia en vivo para 1–8 MP de entrada de referencia.
### Imagen: FLUX 3 Image
@@ -149,14 +154,14 @@ Créditos totales = **(créditos / seg) × `duration`** (`duration` es de 5 a 20
### Video: FLUX 3 Max
-Créditos totales = **(créditos / seg) × `duration`** (`duration` es de 5 a 20 segundos). `reasoning_level` y `generate_audio` no cambian el precio. Las tarifas promocionales de lanzamiento se aplican hasta el **7 de noviembre de 2026 (00:00 UTC)**; después de eso se aplican las tarifas estándar.
+Créditos totales = **(créditos / seg) × `duration`** (`duration` es de 5 a 20 segundos). `reasoning_level` y `generate_audio` no cambian el precio. La tarifa de lanzamiento de 12.82 créditos por segundo en 720p y 21.88 en 1080p se aplicó hasta el 7 de noviembre de 2026 (00:00 UTC). Las filas siguientes muestran las tarifas estándar, que se aplican desde esa fecha.
-| Nodo | Resolución | Créditos promocionales / seg | Créditos estándar / seg |
-| :------------------------ | :--------- | :------------------ | :--------------------- |
-| FLUX 3 Max Text to Video | 720p | 12.82 | 51.29 |
-| FLUX 3 Max Text to Video | 1080p | 21.88 | 87.50 |
-| FLUX 3 Max Image to Video | 720p | 12.82 | 51.29 |
-| FLUX 3 Max Image to Video | 1080p | 21.88 | 87.50 |
+| Nodo | Resolución | Créditos / seg |
+| :------------------------ | :--------- | :------------- |
+| FLUX 3 Max Text to Video | 720p | 51.29 |
+| FLUX 3 Max Text to Video | 1080p | 87.50 |
+| FLUX 3 Max Image to Video | 720p | 51.29 |
+| FLUX 3 Max Image to Video | 1080p | 87.50 |
### Video: Flux Video Edit
@@ -172,10 +177,10 @@ Las tarifas son **por megapíxel de salida por segundo**. Créditos totales = **
| Nodo | Modo | Créditos / MP / seg |
| :----------------- | :-------- | :----------------- |
-| Flux Video Upscale | creative | 15.09 |
-| Flux Video Upscale | precise | 10.56 |
+| Flux Video Upscale | creative | 14.39 |
+| Flux Video Upscale | precise | 10.07 |
-Una salida de 1080p (2.07 MP) cuesta 31.24 créditos por segundo en modo `creative` y 21.86 en `precise`. Una salida de 4K (8.29 MP) cuesta 125.10 y 87.54.
+Una salida de 1080p (2.07 MP) cuesta 29.83 créditos por segundo en modo `creative` y 20.88 en `precise`. Una salida de 4K (8.29 MP) cuesta 119.34 y 83.54.
## Bria
@@ -212,6 +217,18 @@ Una salida de 1080p (2.07 MP) cuesta 31.24 créditos por segundo en modo `creati
| Bria Remove Video Background (Transparent) | 10.55 / seg |
| Bria Video Eraser | 13.58 / seg |
+## Breeze
+
+Todos los nodos de voz de Breeze se cobran a **9.05 créditos por 1K caracteres facturables**, y los caracteres CJK cuentan como dos. Text to Speech y Voice Design cobran el texto que introduces. Voice Clone, Remix y Localize usan la misma tarifa, pero el recuento de caracteres solo se conoce después de la ejecución: Remix y Localize a partir del texto de vista previa generado, y Clone a partir de las unidades que informa el proveedor. Por eso la insignia muestra un intervalo aproximado de **0.66–2.66**, **0.72–1.52** y **0.76–2.29** créditos para esos tres nodos.
+
+| Nodo | Créditos |
+| :-------------------- | :------------------- |
+| Breeze Text to Speech | 9.05 / 1K caracteres |
+| Breeze Voice Design | 9.05 / 1K caracteres |
+| Breeze Voice Clone | 9.05 / 1K caracteres |
+| Breeze Voice Remix | 9.05 / 1K caracteres |
+| Breeze Voice Localize | 9.05 / 1K caracteres |
+
## ByteDance
### LLM: ByteDance Seed
@@ -232,15 +249,17 @@ Por cada imagen generada con éxito.
| modelo | Créditos / imagen |
| :-------------------------------------------------- | :------------------- |
+| seedream 5.5 (`seedream-5-5-261028`) | 6.03 (1K) / 10.56 (2K) |
| seedream 5.0 pro (`seedream-5-0-pro-260628`) | 9.50 (1K) / 18.99 (2K) |
| seedream 5.0 flash (`seedream-5-0-flash-260915`) | 5.43 |
| seedream 5.0 lite (`seedream-5-0-260128`) | 7.39 |
| seedream-4-5-251128 | 8.44 |
| seedream-4-0-250828 | 6.33 |
+| Seedream 5.5 Layer Separation | 6.03 (1K, 1.5K) / 10.56 (2K) |
| Seedream 5.0 Pro Layer Separation | 6.79 (1K, 1.5K) / 13.58 (2K) |
| Seedream 5.0 Flash Layer Separation | 5.43 |
-Seedream 5.0 Pro usa facturación según la resolución: **1K** (≤2.61 MP, **9.50 créditos**) y **2K** (>2.61 MP, **18.99 créditos**). Seedream 5.0 Flash cobra **5.43 créditos** fijos por imagen en cualquier resolución. La primera imagen de cada solicitud es gratis. Cada imagen adicional después de la primera (incluidas tanto las imágenes de salida como las de entrada/referencia) añade **0.63 créditos**.
+Seedream 5.0 Pro usa facturación según la resolución: **1K** (≤2.61 MP, **9.50 créditos**) y **2K** (>2.61 MP, **18.99 créditos**). Seedream 5.0 Flash cobra **5.43 créditos** fijos por imagen en cualquier resolución. La primera imagen de cada solicitud es gratis. Cada imagen adicional después de la primera (incluidas tanto las imágenes de salida como las de entrada/referencia) añade **0.63 créditos**. **Seedream 5.5** cuesta **6.03 créditos** hasta unos 2.6 MP y **10.56 créditos** por encima, y añade **2.41 créditos** por cada imagen de referencia conectada. **Seedream 5.5 Layer Separation** cuesta **6.03 créditos** en 1K/1.5K y **10.56 créditos** en 2K, por las imágenes de la ejecución.
### Imagen: ByteDance Image (obsoleto)
@@ -254,24 +273,44 @@ Seedream 5.0 Pro usa facturación según la resolución: **1K** (≤2.61 MP, **9
### Video: Seedance 2.5
-Facturación basada en tokens. Consumo estimado de tokens = **(duración del video de entrada + duración del video de salida) × ancho del video de salida × alto del video de salida × frecuencia de cuadros del video de salida ÷ 1024**. Cuando la entrada es solo texto, imagen o audio, la duración del video de entrada = 0. El uso real que reporta la API es el valor que prevalece.
+Facturación basada en tokens. La insignia de precio en vivo de ComfyUI estima los tokens como **floor((ancho de salida × alto de salida ÷ 1024) × (24 × duración facturada + 1))** y luego multiplica por la tarifa de Créditos / 1K tokens de la tabla siguiente (el `+ 1` es un fotograma extra). La duración facturada es la de salida, más la del video de entrada cuando hay un video de referencia conectado. Cuando la entrada es solo texto, imagen o audio, la duración del video de entrada = 0. El uso real que reporta la API es el valor que prevalece.
+
+| Modelo | Resolución | Entrada de video | Créditos / 1K tokens |
+| :----------------- | :--------- | :--------------- | -------------------: |
+| Seedance 2.5 | 480p, 720p | no | 3.229 |
+| Seedance 2.5 | 480p, 720p | sí | 1.931 |
+| Seedance 2.5 | 1080p | no | 3.530 |
+| Seedance 2.5 | 1080p | sí | 2.112 |
+| Seedance 2.5 | 4k | no | 1.418 |
+| Seedance 2.5 | 4k | sí | 0.875 |
+| Seedance 2.5 Flash | 480p, 720p | no | 2.112 |
+| Seedance 2.5 Flash | 480p, 720p | sí | 1.297 |
+
+**Entrada de video** significa que hay un video de referencia conectado. Las referencias de texto, imagen y audio cuentan como entrada sin video. La insignia en vivo no aplica un mínimo de tokens adicional más allá de la fórmula anterior.
+
+**Seedance 2.5 Flash** genera solo en 480p y 720p, hasta 30 segundos, y cobra la misma tarifa en ambas. Un clip de 5 segundos en 720p sin entrada de video cuesta unos 230 créditos.
+
+Un fotograma 4k contiene unos cuatro veces más tokens que uno de 1080p con la misma duración, así que el precio por 1K tokens es menor pero el cargo total es mayor.
-| Modelo | Resolución | Entrada de video | Créditos / 1K tokens |
-| :----------- | :--------- | :--------------- | -------------------: |
-| Seedance 2.5 | 480p, 720p | no | 3.229 |
-| Seedance 2.5 | 480p, 720p | sí | 1.931 |
-| Seedance 2.5 | 1080p | no | 3.530 |
-| Seedance 2.5 | 1080p | sí | 2.112 |
+Los videos de referencia deben durar al menos 2 segundos y la entrada total está limitada a 30 segundos. Cargos estimados para una salida en 16:9 a 24 fps, con la fórmula de tokens anterior:
-**Entrada de video** significa que hay un video de referencia conectado. Las referencias de texto, imagen y audio cuentan como entrada sin video. Hay límites mínimos de consumo de tokens cuando la entrada incluye video: si el consumo estimado es inferior al límite, la facturación aplica el mínimo.
+| Resolución | Salida de 5s, sin entrada de video | Salida de 5s + 2s de entrada de video | Salida de 5s + 30s de entrada de video |
+| :--------- | ---------------------------------: | ------------------------------------: | -------------------------------------: |
+| 480p | 156.38 | 130.64 | 650.12 |
+| 720p | 351.58 | 293.72 | 1461.63 |
+| 1080p | 865.00 | 722.82 | 3596.98 |
+| 4k | 1389.91 | 1197.81 | 5960.70 |
-Los videos de referencia deben durar al menos 2 segundos y la entrada total está limitada a 30 segundos. Cargos estimados para una salida en 16:9 a 24 fps:
+Un video de referencia conectado usa la tarifa por 1K tokens más baja, así que 5 segundos más una referencia de 2 segundos pueden costar menos que 5 segundos sin entrada de video. La insignia en vivo no puede ver la duración del video de entrada, así que en una ejecución de 5 segundos con video adjunto el rango va desde solo la duración de salida (a la tarifa con video) hasta la salida más 30 segundos.
-| Resolución | Salida de 5s, sin entrada de video | Salida de 5s + entrada de video de 2–30s |
-| :--------- | ---------------------------------: | ---------------------------------------: |
-| 480p | 155.1 | 166.9 – 649.3 |
-| 720p | 348.8 | 375.4 – 1459.8 |
-| 1080p | 857.8 | 923.9 – 3592.7 |
+Una ejecución con el modelo **Seedance 2.5 Draft** genera una vista previa en 480p (facturada a las tarifas de tokens 480p de la primera tabla anterior) y devuelve un `draft_task_id`. **ByteDance Seedance 2.5 Draft to Final Video** renderiza el vídeo final en 1080p o 4k de ese borrador; el vídeo final reutiliza el prompt, las referencias, la duración y el ajuste de audio del borrador. La insignia en vivo del nodo indica un rango de créditos independiente de la duración, calculado con la misma fórmula (desde un final 16:9 de 4 segundos sin video de referencia, hasta un final de 30 segundos más 30 segundos de video de referencia). El extremo alto es 21:9 en 1080p y 4:3 o 3:4 en 4k:
+
+| Nodo | Resolución | Créditos |
+| :------------------------------------------ | :--------- | ----------------: |
+| ByteDance Seedance 2.5 Draft to Final Video | 1080p | 693.43 – 6202.64 |
+| ByteDance Seedance 2.5 Draft to Final Video | 4k | 1114.23 – 10214.09 |
+
+Un vídeo final de 4 segundos en 1080p 16:9 sin video de referencia cuesta 693.43 créditos, y uno de 30 segundos en 1080p 16:9 con 30 segundos de video de referencia cuesta 6163.19. Ambos totales incluyen el fotograma extra de la insignia. Una proporción distinta cuesta algo más: el máximo de 6202.64 en 1080p corresponde al caso 21:9, y el máximo de 10214.09 en 4k corresponde al caso 4:3 o 3:4. El uso real de la API determina el cargo final.
### Video: Seedance 2.0
@@ -312,7 +351,7 @@ Se factura por tokens (misma fórmula de estimación que Seedance 2.0). A difere
| seedance-1-0-pro-fast-251015 | N/A | 211 |
- **Obsoleto**: `seedance-1-5-pro-251215` se retira el 11 de noviembre de 2026, cuando BytePlus cierre el modelo. ComfyUI lo elimina de los nodos ByteDance 1.x ese mismo día y sus filas de precios ya no aparecen en esta página. Usa `seedance-1-0-pro-250528`, `seedance-1-0-pro-fast-251015` o un modelo Seedance 2.x.
+ **Retirado**: BytePlus cerró `seedance-1-5-pro-251215` el 11 de noviembre de 2026, así que sus filas de precios ya no aparecen en esta página. Usa `seedance-1-0-pro-250528`, `seedance-1-0-pro-fast-251015` o un modelo Seedance 2.x.
**Ya retirado**: ComfyUI retiró y eliminó `seedance-1-0-lite-t2v-250428` y `seedance-1-0-lite-i2v-250428`, además del nodo **ByteDance Reference Images to Video**. Los precios siguientes son solo de referencia.
@@ -404,7 +443,7 @@ A un máximo de 30 fps, una salida en 1080p cuesta unos 62.56 créditos por segu
## FishAudio
-**Oferta por tiempo limitado**: del 13 de octubre al 12 de noviembre de 2026, Drama 3 se factura a aproximadamente **3.02** créditos por 1K bytes. La fila `drama-3` de la tabla muestra el precio estándar, que vuelve a aplicarse a partir del 13 de noviembre de 2026.
+La tarifa de lanzamiento de Drama 3, de aproximadamente **3.02** créditos por 1K bytes, se aplicó del 13 de octubre al 12 de noviembre de 2026. La fila `drama-3` de la tabla muestra el precio estándar, que se aplica desde el 13 de noviembre de 2026.
| Nodo | Modelo | Créditos |
@@ -447,6 +486,9 @@ Facturación basada en tokens; los tokens de pensamiento usan la tarifa del text
| Modelo | Créditos de entrada / 1K | Créditos de salida / 1K |
| :------------------------------------------------------ | -----------------: | :------------------ |
+| Gemini 4 Hydrogen (`gemini-4-hydrogen`) | 0.4526 | 2.263 |
+| Gemini 4 Argon (`gemini-4-argon`) | 1.20692 | 6.0346 |
+| Gemini 4 Carbon (`gemini-4-carbon`) | 1.20692 | 6.0346 |
| Gemini 3.8 Flash (`gemini-3.8-flash`) | 0.4526 | 2.263 |
| Gemini 3.7 Flash (`gemini-3.7-flash`) | 0.4526 | 2.263 |
| Gemini 3.5 Flash (`gemini-3.5-flash`) | 0.3165 | 1.899 |
@@ -478,6 +520,10 @@ Facturación por segundo de video generado. Las tarifas son aproximadas.
| | 720p | 30.57 |
| | 1080p | 45.87 |
| | 4k | 91.76 |
+| Gemini Omni Flash 1.2 (`gemini-omni-1.2-flash`) | 360p | 6.98 |
+| | 720p | 20.97 |
+| | 1080p | 31.46 |
+| | 4k | 62.92 |
### Video: Google Veo
@@ -533,23 +579,33 @@ Los precios se fijan por segundo, por ejecución o por 1K caracteres.
| HeyGen Video Translate | modo velocidad | 4.07 / sec |
| HeyGen Text to Speech | N/A | ~0.20 / sec |
| HeyGen Video 1.0 Reference to Video | 480p, sin referencias | 3.02 / sec |
-| HeyGen Video 1.0 Reference to Video | 768p, sin referencias | 4.53 / sec |
-| HeyGen Video 1.0 Reference to Video | 2K, sin referencias | 13.58 / sec |
-| HeyGen Video 1.0 Reference to Video | 480p, con medios de referencia | 6.03 / sec |
-| HeyGen Video 1.0 Reference to Video | 768p, con medios de referencia | 9.05 / sec |
-| HeyGen Video 1.0 Reference to Video | 2K, con medios de referencia | 27.16 / sec |
+| HeyGen Video 1.0 Reference to Video | 768p, sin referencias | 5.43 / sec |
+| HeyGen Video 1.0 Reference to Video | 1080p, sin referencias | 12.07 / sec |
+| HeyGen Video 1.0 Reference to Video | 2K, sin referencias | 15.09 / sec |
+| HeyGen Video 1.0 Reference to Video | 480p, con medios de referencia | 10.56 / sec |
+| HeyGen Video 1.0 Reference to Video | 768p, con medios de referencia | 15.09 / sec |
+| HeyGen Video 1.0 Reference to Video | 1080p o 2K, con medios de referencia | 27.16 / sec |
+| HeyGen Video 1.0 Reference to Video | imagen extra a partir de la tercera | 6.03 / imagen |
| HeyGen Video 1.0 Image to Video | 480p | 3.02 / sec |
-| HeyGen Video 1.0 Image to Video | 768p | 4.53 / sec |
-| HeyGen Video 1.0 Image to Video | 2K | 13.58 / sec |
+| HeyGen Video 1.0 Image to Video | 768p | 5.43 / sec |
+| HeyGen Video 1.0 Image to Video | 1080p | 12.07 / sec |
+| HeyGen Video 1.0 Image to Video | 2K | 15.09 / sec |
+| HeyGen Video 1.0 Image to Video | más de 15 segundos, 480p | 13.58 / sec |
+| HeyGen Video 1.0 Image to Video | más de 15 segundos, 768p | 22.63 / sec |
+| HeyGen Video 1.0 Image to Video | más de 15 segundos, 1080p o 2K | 31.68 / sec |
| HeyGen Voice 1.0 Instant Clone | N/A | Gratis |
| HeyGen Voice 1.0 Text to Speech | `heygen-voice-1` | 9.05 / 1K chars |
| HeyGen Voice 1.0 Text to Speech | `heygen-voice-1-turbo` | 2.26 / 1K chars |
-**Precios de texto a voz**: hasta el 31 de octubre de 2026 se aplicó una tarifa de lanzamiento de aproximadamente **4.53** créditos por 1K caracteres para `heygen-voice-1` y **1.21** para `heygen-voice-1-turbo`. Las filas anteriores muestran las tarifas estándar, que se aplican a partir del 1 de noviembre de 2026.
+**Precios de texto a voz**: hasta el 1 de noviembre de 2026 (00:00 UTC) se aplicó una tarifa de lanzamiento de aproximadamente **4.53** créditos por 1K caracteres para `heygen-voice-1` y **1.21** para `heygen-voice-1-turbo`. Las filas anteriores muestran las tarifas estándar, que se aplican a partir de ese momento.
El Video de Avatar HeyGen se factura por segundo según la tarifa del tipo de aspecto de avatar seleccionado (Digital Twin, Photo Avatar o Studio Avatar). **Auto** elige el motor para cada aspecto, así que su tarifa cubre todo el rango anterior; un ID de avatar personalizado se factura según su propio tipo.
-Los clips de HeyGen Video 1.0 duran de 5 a 15 segundos y se facturan a la tarifa de la resolución seleccionada. Reference to Video usa la tarifa sin referencias hasta que se conecta una imagen o un video de referencia, y cada video de referencia suma 5 segundos a la duración facturada. La salida en **2K** requiere la relación de aspecto `16:9` o `9:16`; `auto` funciona en 2K solo cuando no hay medios de referencia conectados.
+**Oferta por tiempo limitado**: la tarifa promocional del 50% se aplica hasta el 1 de diciembre de 2026 (00:00 UTC). Las filas de Video 1.0 de la tabla muestran estas tarifas promocionales actuales, salvo las de clips largos, que ya están a precio estándar. Desde el 1 de diciembre de 2026 (00:00 UTC) se aplican las tarifas estándar: sin referencias e Image to Video (≤15 segundos), **6.03 / 10.86 / 24.14 / 30.17** créditos por segundo; con medios de referencia, **21.12 / 30.17 / 54.31** créditos por segundo; cada imagen extra a partir de la tercera, **12.07 créditos**. Las tarifas de clips largos se mantienen.
+
+Los clips de HeyGen Video 1.0 duran de 5 a 30 segundos y se facturan por segundo. Los clips de más de 15 segundos usan las filas de clips largos en ambos nodos cuando no hay medios de referencia. Con referencias conectadas no están disponibles duraciones de más de 15 segundos. Al conectar cualquier imagen o video, **Reference to Video** pasa a la tarifa con referencias, aunque las dos primeras imágenes no tengan cargo por imagen. Las dos primeras imágenes son gratis; cada imagen a partir de la tercera suma **6.03 créditos** (ese recargo pasa a **12.07 créditos** el 1 de diciembre de 2026, 00:00 UTC). Hasta 15 segundos de metraje de referencia se facturan a la tarifa por segundo con referencias.
+
+`auto_duration` sustituye `duration` por la duración del diálogo citado: de 8 a 30 segundos, o de 8 a 15 segundos más el metraje de referencia. La salida en **1080p** y **2K** requiere la relación de aspecto `16:9` o `9:16`; `auto` funciona en esas resoluciones solo sin medios de referencia.
## Higgsfield
@@ -660,15 +716,15 @@ Créditos totales = **(créditos / s) × duración del video**. Las tarifas depe
## Ideogram
-| Nodo | turbo | Referencia de personaje | rendering_speed | Créditos / imagen |
-| :------------- | :---- | :---------------------- | :-------------- | :---------------- |
-| Ideogram V1 | falso | — | — | 18.1 |
-| Ideogram V1 | verdadero | — | — | 6.03 |
-| Ideogram V2 | falso | — | — | 24.14 |
-| Ideogram V2 | verdadero | — | — | 15.09 |
-| Ideogram V4 | — | — | TURBO | 9.05 |
-| Ideogram V4 | — | — | DEFAULT | 18.1 |
-| Ideogram V4 | — | — | QUALITY | 30.17 |
+
+ **Obsoleto**: ComfyUI eliminó los nodos **Ideogram V1** e **Ideogram V2**, así que esta página ya no muestra sus precios. Usa **Ideogram V4** en su lugar.
+
+
+| Nodo | rendering_speed | Créditos / imagen |
+| :---------- | :-------------- | :---------------: |
+| Ideogram V4 | TURBO | 9.05 |
+| Ideogram V4 | DEFAULT | 18.1 |
+| Ideogram V4 | QUALITY | 30.17 |
| Nodo | Calidad | Resolución | Créditos / imagen |
| :---------------- | :-------- | :--------- | :---------------- |
@@ -727,13 +783,41 @@ Créditos totales = **créditos / imagen** con la `resolution` y el `thinking_mo
| Krea 3 | 2K | high | 24.14 | 6.03 - 21.12 |
| Krea 3 | 2K | ultra | 78.45 | 6.03 - 21.12 |
-A 2K, cada imagen de referencia añade **de 6.03 a 21.12 créditos**, según el tamaño en píxeles de la imagen de referencia; a 1K, las imágenes de referencia se reducen a aproximadamente 1 MP, por lo que el recargo es fijo de **6.03 créditos**. `ultra` siempre genera 2K, independientemente de `resolution`, y requiere al menos una imagen de referencia.
+A 2K, cada imagen de referencia añade **6.03–21.12 créditos** (**6.03** en el extremo inferior y **21.12** en el superior; el nodo escala entre estos dos valores según el tamaño en píxeles de la imagen de referencia). A 1K, las imágenes de referencia se reducen a aproximadamente 1 MP, por lo que el recargo es fijo de **6.03 créditos**. `ultra` siempre genera 2K, independientemente de `resolution`, y requiere al menos una imagen de referencia.
## Kling
### Video: créditos / seg
-Créditos totales = **(créditos / seg) × `duration`**, excepto **Kling 3.0 Omni Edit Video** y **Kling Motion Control** (con cobro por segundo de salida; la duración sigue al video de entrada en Edit).
+Créditos totales = **(créditos / seg) × `duration`**, excepto **Kling 3.0 Omni Edit Video**, **Kling 4.0 Edit Video** y **Kling Motion Control** (con cobro por segundo de salida; la duración sigue al video de entrada en Edit).
+
+#### kling-4.0
+
+`kling-4.0` genera de 3 a 30 segundos en 720p o 1080p; `kling-4.0-flash` genera de 3 a 20 segundos en 720p.
+
+| Nodo | Modelo | Resolución | Créditos / seg | Créditos / seg con video de referencia |
+| :--------------------------- | :-------------- | :--------- | :------------- | :------------------------------------- |
+| Kling 4.0 Text to Video | kling-4.0 | 720p | 47.37 | — |
+| Kling 4.0 Text to Video | kling-4.0 | 1080p | 70.91 | — |
+| Kling 4.0 Image to Video | kling-4.0 | 720p | 47.37 | — |
+| Kling 4.0 Image to Video | kling-4.0 | 1080p | 70.91 | — |
+| Kling 4.0 Reference to Video | kling-4.0 | 720p | 47.37 | 28.36 |
+| Kling 4.0 Reference to Video | kling-4.0 | 1080p | 70.91 | 42.54 |
+| Kling 4.0 Text to Video | kling-4.0-flash | 720p | 28.36 | — |
+| Kling 4.0 Image to Video | kling-4.0-flash | 720p | 28.36 | — |
+| Kling 4.0 Reference to Video | kling-4.0-flash | 720p | 28.36 | 19.01 |
+
+En **Kling 4.0 Reference to Video**, los videos de referencia conectados cambian toda la ejecución a la tarifa más baja: cada segundo de salida y cada segundo de referencia se facturan a esa tarifa. Cada clip dura de 3 a 30 segundos (de 3 a 20 en `kling-4.0-flash`) y todos juntos deben caber en ese mismo límite. Una ejecución de 5 segundos en 720p con 30 segundos de video de referencia cuesta 992.7 créditos.
+
+**Kling 4.0 Edit Video** siempre cobra a la tarifa más baja (la de video de referencia): **créditos = tarifa × (segundos de video de entrada + segundos de video de referencia)**. La salida tiene la misma duración que la entrada, así que sin videos de referencia es **2 × la tarifa** por segundo de entrada. Con videos de referencia, el cargo real sigue esa misma fórmula; la insignia en vivo muestra de **1× a 2× la tarifa** por segundo de entrada porque no puede ver las duraciones. Las imágenes de referencia no añaden un cargo por imagen.
+
+| Modelo | Resolución | Tarifa (créditos / seg) | Sin videos de referencia | Con videos de referencia |
+| :-------------- | :--------- | ----------------------: | :----------------------- | :----------------------- |
+| kling-4.0 | 720p | 28.36 | 56.73 / seg de entrada (2 × tarifa) | 28.36 × (entrada + segundos de ref.) |
+| kling-4.0 | 1080p | 42.54 | 85.09 / seg de entrada (2 × tarifa) | 42.54 × (entrada + segundos de ref.) |
+| kling-4.0-flash | 720p | 19.01 | 38.02 / seg de entrada (2 × tarifa) | 19.01 × (entrada + segundos de ref.) |
+
+Una edición de 10 segundos en 1080p sin videos de referencia cuesta 850.9 créditos.
#### kling-video-o1
@@ -789,10 +873,6 @@ Siempre genera audio; el interruptor de audio no tiene efecto.
| motion control | — | std | 14.77 |
| motion control | — | pro | 23.63 |
-
- **Fin de vida útil (EOL)**: Kling V1.5, V1.6, V2.1, V2.1 Master y Kolors Virtual Try-On se retirarán el **15 de septiembre de 2026**. La tarifa a continuación se conserva como referencia hasta su eliminación.
-
-
#### kling-v2-5-turbo
**Kling Text to Video**, **Kling Image (First Frame) to Video** y **Kling Start-End Frame to Video** usan `kling-v2-5-turbo` solo en modo **pro**, con cobro por ejecución a una duración fija de 5s o 10s.
@@ -840,6 +920,10 @@ Por ejecución.
| :------------------- | :------------------- |
| Kling Virtual Try-On | 33.79 |
+
+ **Retirado**: Kling retiró V1.5, V1.6, V2.1, V2.1 Master y el servicio anterior Kolors Virtual Try-On el **15 de septiembre de 2026**. Los nodos de Kling ya no ofrecen esos modelos, por lo que sus filas de precios se han eliminado de esta página. La fila **Kling Virtual Try-On** de arriba es otro nodo vigente que se factura por ejecución.
+
+
## Lightricks
### Video: créditos / seg
@@ -869,15 +953,35 @@ Créditos totales = **(créditos / seg) × `duration`**.
La resolución corresponde a `1280x720` / `720x1280` (720p), `1920x1080` / `1080x1920` (1080p), `2560x1440` / `1440x2560` (1440p) y `3840x2160` / `2160x3840` (4K). Las duraciones superiores a 10 s en **LTX-2.5 (Fast)** requieren una resolución de 720p o 1080p y 24/25 FPS; en **LTX-2 (Fast)** requieren 1080p y 25 FPS.
+### Posprocesamiento: créditos / seg
+
+Estos nodos procesan el vídeo de entrada y cobran por segundo del mismo.
+
+| Nodo | resolución | Créditos / seg |
+| :------------------- | ---------- | --------------- |
+| LTX Video AlphaGen | 1080p (lado corto ≤1080; por debajo de 1080 usa este tramo) | 30.17 |
+| LTX Video AlphaGen | 1440p (lado corto ≤1440) | 75.43 |
+| LTX Video AlphaGen | 2160p (lado corto ≤2160) | 301.73 |
+| LTX Video SDR to HDR | 1080p (≤ 1920×1080) | 120.69 |
+| LTX Video SDR to HDR | 1440p (≤ 2560×1440) | 241.39 |
+| LTX Video SDR to HDR | 4K (≤ 3840×2160) | 482.77 |
+| LTX Video Reframe | 720p | 30.17 |
+| LTX Video Reframe | 1080p | 60.35 |
+| LTX Video Upscale | 1080p | 24.14 |
+| LTX Video Upscale | 4k | 90.52 |
+| LTX Video Upscale | 8k | 362.08 |
+
+**AlphaGen** se cobra por segundo de vídeo de entrada, según el lado corto de la entrada (el menor entre el ancho y el alto). Las entradas con un lado corto inferior a 1080 píxeles usan la tarifa de 1080p. **SDR to HDR** también se cobra por segundo de vídeo de entrada; la entrada se ajusta al tramo de resolución más pequeño que pueda contener su recuento de píxeles. La insignia en vivo de ComfyUI no puede ver la resolución de entrada, así que solo muestra el intervalo (**30.17–301.73** créditos/s para AlphaGen, **120.69–482.77** para SDR to HDR). El cargo real usa las tres tarifas de la tabla. **AlphaGen** admite hasta 10 segundos de vídeo, **Reframe** hasta 60, y **SDR to HDR** y **Upscale** limitan el número de fotogramas según la resolución (273 fotogramas en 1080p para SDR to HDR, 2,401 para Upscale).
+
## Luma
### Video
- **Obsoleto**: Los nodos clásicos **Luma Text to Video** y **Luma Image to Video** están obsoletos, y Luma eliminó los endpoints que usan el 24 de octubre de 2026, así que estos dos nodos ya no pueden generar videos. Usa los nodos **Luma Ray 3.2** en su lugar. Las filas de ray-2, ray-flash-2 y ray-1-6 a continuación se mantienen solo como referencia; **ray-1-6** dejó de servirse en septiembre de 2026.
+ **Retirado**: los nodos clásicos Luma Text to Video, Image to Video y de generación de imágenes ya no existen, y los modelos ray-2, ray-flash-2, ray-1-6, photon-1 y photon-flash-1 tampoco se ofrecen. Las tablas siguientes solo incluyen los nodos que ComfyUI sigue ofreciendo.
-Los precios fijos de **/run** coinciden con la insignia del nodo para cada opción de `duration`. **Edit** y **reframe** cobran **(créditos / seg) × duración del video**. **ray-1-6** siempre cuesta **105.50 / run**.
+Los precios fijos de **/run** coinciden con la insignia del nodo para cada opción de `duration`. **Edit** y **reframe** cobran **(créditos / seg) × duración del video**.
| Modelo | tarea | resolución | Créditos |
| :---------- | :--------- | :--------- | :------------------------------------------------ |
@@ -893,31 +997,18 @@ Los precios fijos de **/run** coinciden con la insignia del nodo para cada opci
| ray-3.2 | reframe | 540p | 12.66 / sec |
| ray-3.2 | reframe | 720p | 25.32 / sec |
| ray-3.2 | reframe | 1080p | 75.96 / sec |
-| ray-2 | generation | 540p | 120.27 / run (5s), 217.33 / run (9s) |
-| ray-2 | generation | 720p | 215.22 / run (5s), 386.13 / run (9s) |
-| ray-2 | generation | 1080p | 478.97 / run (5s), 865.10 / run (9s) |
-| ray-2 | generation | 4K | 1922.21 / run (5s), 3460.40 / run (9s) |
-| ray-flash-2 | generation | 540p | 42.20 / run (5s), 75.96 / run (9s) |
-| ray-flash-2 | generation | 720p | 71.74 / run (5s), 128.71 / run (9s) |
-| ray-flash-2 | generation | 1080p | 166.69 / run (5s), 299.62 / run (9s) |
-| ray-flash-2 | generation | 4K | 660.43 / run (5s), 1192.15 / run (9s) |
-| ray-1-6 | generation | — | 105.50 / run |
-
-**Luma Ray 3.2** Image to Video y Extend Video siempre cobran el precio de **5s**. El clásico Image to Video usa **5s** o **9s** de las filas de **ray-2** / **ray-flash-2**.
+
+**Luma Ray 3.2** Image to Video y Extend Video siempre cobran el precio de **5s**.
### Imagen
| Modelo | tarea | Créditos |
| :------------- | :--------- | :----------------------------------- |
-| photon-flash-1 | — | 0.57 / run |
-| photon-1 | — | 2.19 / run |
| uni-1 | generation | 8.52 / run + 0.63 / ref image (≤ 9) |
| uni-1 | edit | 9.16 / run + 0.63 / ref image (≤ 8) |
| uni-1-max | generation | 21.10 / run + 0.63 / ref image (≤ 9) |
| uni-1-max | edit | 21.73 / run + 0.63 / ref image (≤ 8) |
-El recargo por imagen de referencia usa USD redondeado a 4 decimales antes de convertirlo a créditos.
-
## Meshy
| Nodo | Opciones | Créditos |
@@ -976,7 +1067,7 @@ Créditos totales = **(créditos / seg) × segundos facturados**. **Mirelo Video
| MiniMax H3 First-Last-Frame to Video | MiniMax H3 | 2K | 39.22 / s |
| MiniMax H3 Reference to Video | MiniMax H3 | 768P | 27.16 / s, más 12.07 / imagen a partir de 5 |
| MiniMax H3 Reference to Video | MiniMax H3 | 2K | 39.22 / s, más 12.07 / imagen a partir de 5 |
-| MiniMax H3 Context IR | MiniMax H3 | — | 4.22 / run (solo texto), 10.55 / run (con primer/último fotograma), 12.66–23.21 / run (con medios de referencia) |
+| MiniMax H3 Context IR | MiniMax H3 | — | 271.56 / 1M tokens de entrada + 1,086.23 / 1M tokens de salida |
| MiniMax H3 Regenerate to 2K | MiniMax H3 | — | 15.09 / s |
| MiniMax H3 Text to Video | MiniMax H3 Max | 480P | 15.09 / s |
| MiniMax H3 Text to Video | MiniMax H3 Max | 768P | 24.14 / s |
@@ -1009,7 +1100,7 @@ Créditos totales = **(créditos / seg) × segundos facturados**. **Mirelo Video
**MiniMax Hailuo 02 Video** solo admite **6s** en 1080P.
-**MiniMax H3** cobra por segundo (duración de 4 a 15s). **MiniMax H3 Reference to Video** suma 12.07 créditos por cada imagen de referencia a partir de la quinta y cobra los videos de referencia a 27.16 créditos/s (768P) o 39.22 créditos/s (2K) (2s por video, hasta 15s en total, aproximado). Los precios de **MiniMax H3 Context IR** dependen de los medios adjuntos: solo texto, primer/último fotograma o medios de referencia (aproximado).
+**MiniMax H3** cobra por segundo (duración de 4 a 15s). **MiniMax H3 Reference to Video** suma 12.07 créditos por cada imagen de referencia a partir de la quinta y cobra los videos de referencia a 27.16 créditos/s (768P) o 39.22 créditos/s (2K) (2s por video, hasta 15s en total, aproximado). **MiniMax H3 Context IR** se cobra a **271.56 créditos por 1M tokens de entrada** y **1,086.23 créditos por 1M tokens de salida**. La insignia del nodo no puede ver el recuento de tokens al componer el grafo, así que muestra un intervalo aproximado de unos **4.22** (solo texto), **10.55** (con primer/último fotograma) o **12.66–23.21** (con medios de referencia).
**MiniMax H3 Max** cobra por segundo con las tarifas anteriores. **Text to Video**, **First-Last-Frame to Video** y **Reference to Video** generan salidas de **5 a 15 s**. **Lip Sync** acepta audio de 5 a ~899 segundos; si la salida supera 15 segundos, se aplica el multiplicador **1.2×** indicado más abajo.
@@ -1035,6 +1126,19 @@ Ejemplos completados, cada uno suponiendo que ninguna otra entrada de referencia
**MiniMax H3 Max Recast**, **Extend Video**, **Reframe Video**, **Insert Video**, **Lip Sync** y **Camera Control** cobran por segundo de video. Extend Video también funciona con el modelo Turbo a la mitad de la tarifa Max. **Reframe Video** se cobra a la tarifa de la tabla por segundo. Las fuentes más largas usan ventanas superpuestas, así que los segundos facturados pueden superar la duración final de la salida. **Lip Sync** se cobra a la tarifa de la tabla por segundo. Cuando la salida **supera los 15 segundos**, se aplica un multiplicador de **1.2×** (por ejemplo, 768P → 28.97 / s). Insert Video suma tokens de referencia por encima de la asignación gratuita de 4,096 tokens a 6.03 créditos por 1K tokens, la misma asignación que Reference to Video: 240 tokens base, 2,048 por imagen y 225,000 por video de referencia con salida a 768P (87,000 a 480P).
+### Imagen: MiniMax H3 Image
+
+Por cada imagen generada. El cargo combina el tamaño de salida con el prompt y las imágenes de referencia, que se facturan como tokens.
+
+| Nodo | modelo | tamaño | Créditos / imagen |
+| :--------------- | -------------------------- | ------ | ----------------- |
+| MiniMax H3 Image | MiniMax-H3-Image | 2K | 5.32 – 8.17 |
+| MiniMax H3 Image | MiniMax-H3-Image | 4K | 19.41 – 22.26 |
+| MiniMax H3 Image | MiniMax-H3-Image-Highspeed | 2K | 10.02 – 12.86 |
+| MiniMax H3 Image | MiniMax-H3-Image-Highspeed | 4K | 38.20 – 41.04 |
+
+La tabla es la franja de una solicitud solo de texto con `enhance_prompt` activado. La fórmula completa es: créditos de salida = `max(píxeles, 4,194,304) / 1,024 × 0.00114657` (usar `0.00229315` en Highspeed), es decir **4.70 créditos** en 2K (4 MP) y unos **18.79 créditos** en 4K, o **9.39** / **37.57** en Highspeed. Los tokens de prompt se facturan a **0.000694 créditos** cada uno: **0–300** sin referencias y con la mejora desactivada, **900–5,000** sin referencias y con la mejora activada, o **2,000 + 300 × (n − 1)** a **6,500 + 800 × (n − 1)** con `n` imágenes de referencia. Cada imagen de referencia añade también **2,050–8,194** tokens de imagen a **0.000453 créditos** cada uno (**0.000905** en Highspeed), unos **0.93 a 3.71 créditos** (**1.86 a 7.42** en Highspeed). Una imagen 2K con la mejora desactivada cuesta unos **4.70 créditos**. Un tamaño personalizado factura su propia superficie con un mínimo de 4 MP.
+
## OpenAI
DALL·E y Sora usan tarifas fijas por imagen o por segundo. **OpenAI GPT Image 2** y **OpenAI ChatGPT** facturan según el uso de tokens de la API.
@@ -1085,14 +1189,55 @@ Los modelos **GPT Image 2.5** (`gpt-image-2.5-flare`, `gpt-image-2.5-sunburst`)
Los recuentos de tokens de salida escalan con el tamaño y la calidad (por ejemplo, 1024x1024 con `low` es de unos 196 tokens de salida, unos 1.32 créditos con `gpt-image-1.5`). Cada imagen de referencia de 1024x1024 añade unos 2.47 créditos de entrada de imagen, y el texto del prompt se factura a la tarifa de entrada de texto. Multiplicar las imágenes (`n`) escala el total.
+Los modelos **GPT Image 2.6** (`gpt-image-2.6-flare`, `gpt-image-2.6-sunburst`) comparten una sola tabla de tarifas. Los créditos son por imagen a 1024x1024:
+
+| calidad | Créditos / imagen |
+| :------ | :---------------- |
+| low | 7.45 |
+| medium | 13.12 |
+| high | 15.89 |
+| xhigh | 28.25 |
+| max | 63.57 |
+
+Los tamaños mayores cuestan más: 3840x2160 son 14.12 créditos con `low` y 120.8 con `max`. Créditos totales = **(precio de salida + referencias × 2.47 a 3.71) × `n`**. Cada imagen de referencia añade **2.47–3.71 créditos**. La insignia en vivo no puede ver el tamaño de la referencia que fija el cargo exacto, así que muestra ese intervalo; `n` multiplica el total.
+
+### Imagen: GPT Image 2.6 GIF
+
+**OpenAI GPT Image 2.6 GIF** cobra por ejecución, no por token. Cada fotograma es una imagen completa, así que el tamaño del fotograma, la calidad y `frame_count` fijan el precio. `frame_count` factura por tramos de 2 a 8, 9, 10 a 12 y 13 a 16 fotogramas, y los tamaños mayores solo admiten los tramos bajos. Créditos para una ejecución de 512x512 con 9 fotogramas:
+
+| calidad | Créditos / ejecución |
+| :------ | :------------------- |
+| low | 10.61 |
+| medium | 18.76 |
+| high | 22.72 |
+| xhigh | 40.41 |
+| max | 90.92 |
+
+Una ejecución de 384x384 con 2 a 8 fotogramas cuesta 3.88 créditos con `low`, y una de 720x720 con 13 a 16 fotogramas cuesta 214.69 con `max`. `frame_duration_ms` no cambia el precio.
+
+### Imagen: GPT Image 2.6 SVG
+
+**OpenAI GPT Image 2.6 SVG** cobra el precio de imagen fija del tamaño y la calidad elegidos, más **0.23–1.77 créditos** por el paso vectorial. La insignia en vivo no puede ver la entrada que fija el cargo exacto del paso vectorial, así que muestra ese intervalo sobre el precio de imagen fija. Los rangos de la tabla son ese precio más ese intervalo. La insignia de SVG no añade un cargo aparte por imagen de referencia. Los tamaños son 1024x1024, 1024x1536 y 1536x1024. Créditos a 1024x1024:
+
+| calidad | Créditos / ejecución |
+| :------ | :------------------- |
+| low | 7.68 – 9.22 |
+| medium | 13.36 – 14.90 |
+| high | 16.12 – 17.66 |
+| xhigh | 28.48 – 30.03 |
+| max | 63.81 – 65.35 |
+
+Una ejecución de 1024x1536 o 1536x1024 cuesta menos: de 6.25 a 7.79 créditos con `low`, y de 49.90 a 51.44 con `max`.
+
### Chat
Facturación basada en tokens. Las tarifas de la insignia son **por 1K tokens**; la tabla es **por 1M tokens**.
| Modelo | Créditos de entrada / 1M | Créditos de salida / 1M |
| :----------- | -----------------: | :------------------ |
+| gpt-6.1-sol, gpt-6-sol | 603.46 | 3017.3 |
+| gpt-6.1-terra | 301.73 | 1508.65 |
| gpt-6-astra | 3017.3 | 15086.5 |
-| gpt-6-sol | 603.46 | 3017.3 |
| gpt-6-luna | 30.17 | 150.87 |
| gpt-5.6-sol, gpt-5.6-sol-pro | 1055 | 6330 |
| gpt-5.6-terra, gpt-5.6-terra-pro | 527.5 | 3165 |
@@ -1152,10 +1297,12 @@ Las tarifas de la insignia son **por 1K tokens**; la tabla es **por 1M tokens**.
| x-ai/grok-4.5 | 603.46 | 1810.38 |
| x-ai/grok-4.20 | 377.16 | 754.33 |
| x-ai/grok-4.3 | 377.16 | 754.33 |
+| meta/muse-wonder | 1206.92 | 6034.60 |
+| meta/muse-spark-1.3 | 377.16 | 1282.35 |
+| meta/muse-ember | 30.17 | 150.87 |
| deepseek/deepseek-v4-pro | 131.25 | 262.51 |
| deepseek/deepseek-v4-flash | 33.79 | 67.59 |
| deepseek/deepseek-v3.2 | 76.04 | 114.05 |
-| qwen/qwen3.6-max-preview | 313.80 | 1882.80 |
| qwen/qwen3.6-plus | 98.06 | 588.37 |
| qwen/qwen3.6-flash | 56.57 | 339.45 |
| mistralai/mistral-large-2512 | 150.87 | 452.60 |
@@ -1197,6 +1344,39 @@ Los tres nodos cobran por ejecución y cada mapa PBR que actives suma al precio.
Los precios son aproximados. Un material de 1 MP con los mapas `basecolor` y `normal` por defecto cuesta 15.09 créditos, y una salida de 4 MP con los cinco mapas a 4× cuesta 184.06 créditos con **Patina Material** o 211.21 créditos con **Patina Material Extract**.
+## Pika
+
+Pika H3 Flex se factura por segundo de vídeo. Créditos totales = **(créditos / seg) × segundos facturados**. `duration` va de 5 a 15 segundos. En `480p` y `768p` la duración facturada es la de `duration`. Desde `1080p` el nodo ajusta el clip a su rejilla de fotogramas, así que una generación de 5 segundos factura 5.17 segundos, una de 10 segundos factura 10.13 y una de 15 segundos factura 15.08.
+
+| Nodo | Resolución | Créditos / seg |
+| :------------------------------ | :--------- | :------------- |
+| Pika H3 Flex Video | 480p | 3.02 |
+| Pika H3 Flex Video | 768p | 6.03 |
+| Pika H3 Flex Video | 1080p | 10.56 |
+| Pika H3 Flex Video | 2k | 12.07 |
+| Pika H3 Flex Video | 4k | 24.14 |
+| Pika H3 Flex Reference to Video | 480p | 9.05 |
+| Pika H3 Flex Reference to Video | 768p | 18.10 |
+| Pika H3 Flex Reference to Video | 1080p | 22.63 |
+| Pika H3 Flex Reference to Video | 2k | 24.14 |
+| Pika H3 Flex Reference to Video | 4k | 30.17 |
+
+Una generación de 5 segundos con **Pika H3 Flex Video** en 480p cuesta 15.09 créditos, y en 1080p cuesta 54.56.
+
+**Pika H3 Flex Reference to Video** factura los videos de referencia conectados con la misma tarifa. Cada clip dura de 2 a 15 segundos y el total de video de referencia está limitado a 15 segundos, así que el costo depende de cuánto video de referencia conectes. En 480p, una generación de 5 segundos con una referencia de 2 segundos cuesta 63.36 créditos. En 4k, una generación de 5 segundos con 15 segundos de video de referencia cuesta 608.49. Todos los ejemplos de Pika usan las tarifas por segundo sin redondear.
+
+### Audio: Pika
+
+Los nodos de audio de Pika también se facturan por segundo. **Pika Soundtrack** cobra la duración del video de entrada (hasta 10 minutos), **Pika SFX** cobra `duration` (1 a 20 segundos) y **Pika Music** cobra la duración de la pista: `duration` en `text to music` (10 a 360 segundos), o la duración del audio de origen limitada a 5 a 360 segundos en `cover`.
+
+| Nodo | Créditos / seg |
+| :-------------- | :------------- |
+| Pika Soundtrack | 1.509 |
+| Pika SFX | 0.060 |
+| Pika Music | 0.075 |
+
+Una ejecución de Soundtrack de 10 minutos cuesta 905.2 créditos, una de SFX de 20 segundos 1.21 y una pista de Music de 3 minutos 13.58.
+
## Pixverse
Modelo **v3.5** (fijo en la API; no seleccionable).
@@ -1223,19 +1403,12 @@ Los nodos de PixVerse V6 generan videos de 1 a 15 segundos con audio nativo opci
| PixVerse V6 | 540p | 7.18 | 9.23 |
| PixVerse V6 | 720p | 9.23 | 12.31 |
| PixVerse V6 | 1080p | 18.47 | 23.60 |
-| PixVerse V6 Max | 360p | 6.71 | 9.39 |
-| PixVerse V6 Max | 540p | 9.39 | 12.07 |
-| PixVerse V6 Max | 720p | 12.07 | 16.09 |
-| PixVerse V6 Max | 1080p | 24.14 | 30.84 |
-
-**PixVerse V6 Max** se ejecuta con las tarifas anteriores como precio introductorio. A partir del **3 de noviembre de 2026**, las tarifas por segundo son:
+| PixVerse V6 Max | 360p | 8.05 | 12.07 |
+| PixVerse V6 Max | 540p | 12.07 | 14.75 |
+| PixVerse V6 Max | 720p | 14.75 | 20.12 |
+| PixVerse V6 Max | 1080p | 29.50 | 37.55 |
-| calidad | Audio desactivado | Audio activado |
-| :------ | :---------------- | :------------- |
-| 360p | 8.05 | 12.07 |
-| 540p | 12.07 | 14.75 |
-| 720p | 14.75 | 20.12 |
-| 1080p | 29.50 | 37.55 |
+**PixVerse V6 Max** tuvo tarifas introductorias más bajas hasta el **2 de noviembre de 2026**. Las filas anteriores muestran las tarifas estándar, vigentes desde el **3 de noviembre de 2026**.
**PixVerse V6 Fusion** cobra el doble de la tarifa por segundo por cada segundo de video de referencia, en lugar de la duración de salida. Fusion solo usa las tarifas de `PixVerse V6`.
@@ -1243,21 +1416,10 @@ Los nodos de PixVerse V6 generan videos de 1 a 15 segundos con audio nativo opci
### Imagen
-Los modelos **arrow-1.1**, ya retirados, cobraban una cantidad fija por ejecución:
-
- **Obsoleto**: Quiver retiró los modelos Arrow 1-series el **16 de octubre de 2026**, y los nodos Quiver Text to SVG y Quiver Image to SVG ya no incluyen `arrow-1.1`, `arrow-1.1-max` ni `arrow-preview`. Usa los modelos Arrow 2 en su lugar. El precio que aparece a continuación se conserva solo como referencia.
+ **Retirado**: Quiver retiró los modelos Arrow 1-series (`arrow-1.1`, `arrow-1.1-max`, `arrow-preview`) el **16 de octubre de 2026**, y los nodos Quiver Text to SVG y Quiver Image to SVG ya no los incluyen. Usa los modelos Arrow 2 de abajo.
-| modelo | Nodo | Créditos / run |
-| :------------ | :------------------ | :------------------- |
-| arrow-1.1 | Quiver Text to SVG | 60.35 |
-| arrow-1.1 | Quiver Image to SVG | 45.26 |
-| arrow-1.1-max | Quiver Text to SVG | 75.43 |
-| arrow-1.1-max | Quiver Image to SVG | 60.35 |
-| arrow-preview | Quiver Text to SVG | 90.52 |
-| arrow-preview | Quiver Image to SVG | 90.52 |
-
**arrow-2** y **arrow-2-telos** cobran según los tokens que consume cada solicitud:
| modelo | Créditos de entrada / 1M tokens | Créditos de salida / 1M tokens |
@@ -1378,12 +1540,15 @@ Gen4 Turbo: **5s** o **10s**. Créditos totales = **(créditos / seg) × `durati
## Sonilo
-Texto a Música: créditos totales = **(créditos / seg) × `duration`** (1–360 segundos).
+Video a Música se factura por segundo de vídeo de entrada, hasta 6 minutos. Texto a Música y Music 1.5 se facturan por segundo de audio: créditos totales = **(créditos / seg) × `duration`**.
| Nodo | Créditos |
| :-------------------- | :----------- |
-| Sonilo Video to Music | 1.9 / seg |
-| Sonilo Text to Music | 0.5275 / seg |
+| Sonilo Video to Music | 2.72 / seg |
+| Sonilo Text to Music | 0.6789 / seg |
+| Sonilo Music 1.5 | 0.6789 / seg |
+
+Texto a Música genera una duración de 1 a 360 segundos. Music 1.5 ajusta la canción a la letra con `auto`, o genera la duración `custom` que indiques, de 10 a 360 segundos.
## sync.so
@@ -1454,7 +1619,7 @@ Multiview solo se aplica a **Hunyuan3D: Image(s) to Model** cuando conectas `ima
| :-------- | :----------- |
| Reimagine | 67.52 / 8MP |
| Bloom 2 | 96.55 / 8MP |
-| Wonder 3.5 | 24.14 / 8MP |
+| Wonder 3.5, Wonder 4 | 24.14 / 8MP |
**Topaz Image Enhance (Legacy)** está obsoleto. No tiene una estimación fija antes de la ejecución; el cargo proviene de la estimación de la API al momento de la consulta (créditos de la API × 0.08 USD). Cargos típicos según la resolución de entrada:
@@ -1614,6 +1779,19 @@ Pruna P-Video-2 cobra por segundo de video generado. Créditos totales = **(cré
| | 1080p | off | 15.09 |
| | 1080p | on | 9.05 |
+### Vídeo: P-Video Upscale
+
+Pruna P-Video Upscale cobra por segundo de vídeo de salida. Créditos totales = **(créditos / segundo) × duración de salida**. La tarifa depende de `target`, una salida superior a 30 fps cuesta el doble y un origen mayor que el objetivo conserva su tamaño.
+
+La tarifa de lanzamiento de 1.51, 3.02, 6.03 y 12.07 créditos por segundo para `720p`, `1080p`, `2K` y `4K` se aplicó hasta el 12 de noviembre de 2026 (17:00 UTC). Las filas siguientes muestran las tarifas estándar, que se aplican desde esa fecha.
+
+| target | Tamaño objetivo | Créditos / segundo |
+| :----- | :-------------- | :----------------- |
+| 720p | unos 0.9 MP | 3.02 |
+| 1080p | unos 2.1 MP | 6.03 |
+| 2K | unos 3.7 MP | 12.07 |
+| 4K | unos 8.3 MP | 24.14 |
+
## VEED
| Nodo | Opciones | Créditos |
@@ -1703,23 +1881,25 @@ Duración fija de **5 s** a **1080p**.
| wan2.5-t2i-preview, wan2.5-i2i-preview | - | 6.33 / run |
| wan2.7-t2v, wan2.7-i2v | 720P | 21.1 / seg |
| | 1080P | 31.65 / seg |
-| wan2.7-i2v (continuation) | 720P | 21.1 / seg (alcance) |
-| | 1080P | 31.65 / seg (alcance) |
+| wan2.7-i2v (continuation) | 720P | 21.1 / seg (salida + 2–5s de entrada) |
+| | 1080P | 31.65 / seg (salida + 2–5s de entrada) |
| wan2.7-videoedit | 720P | 21.1 / seg (entrada + salida) |
| | 1080P | 31.65 / seg (entrada + salida) |
-| wan2.7-r2v | 720P | 21.1 / seg (alcance) |
-| | 1080P | 31.65 / seg (alcance) |
-| wan3.0-video | 480P | 15.09 / seg (alcance) |
-| | 720P | 30.17 / seg (alcance) |
-| | 1080P | 60.35 / seg (alcance) |
-| wan3.0-video-prime | 480P | 20.52 / seg (alcance) |
-| | 720P | 42.24 / seg (alcance) |
-| | 1080P | 84.48 / seg (alcance) |
+| wan2.7-r2v | 720P | 21.1 / seg (salida hasta salida + 5s de entrada) |
+| | 1080P | 31.65 / seg (salida hasta salida + 5s de entrada) |
+| wan3.0-video | 480P | 15.09 / seg |
+| | 720P | 30.17 / seg |
+| | 1080P | 60.35 / seg |
+| wan3.0-video-prime | 480P | 20.52 / seg |
+| | 720P | 42.24 / seg |
+| | 1080P | 84.48 / seg |
| wan2.5-t2v-preview, wan2.5-i2v-preview | 480p | 10.55 / seg |
| wan2.5-t2v-preview, wan2.5-i2v-preview, wan2.6-t2v, wan2.6-i2v | 720p | 21.1 / seg |
| wan2.5-t2v-preview, wan2.5-i2v-preview, wan2.6-t2v, wan2.6-i2v | 1080p | 31.65 / seg |
-| wan2.6-r2v | 720p | 21.1 / seg (alcance) |
-| | 1080p | 31.65 / seg (alcance) |
+| wan2.6-r2v | 720p | 21.1 / seg (salida + 2–5s de entrada) |
+| | 1080p | 31.65 / seg (salida + 2–5s de entrada) |
+
+**Wan 2.6 Reference to Video** y **Wan 2.7 continuation** cobran la duración de salida más **2–5 segundos** de entrada, a la misma tarifa por segundo. **Wan 2.7 Reference to Video** va desde solo la duración de salida hasta esa duración más **5 segundos** de entrada. **Wan 2.7 Edit** cobra entrada más salida a la tarifa de la tabla. **Wan 3.0** cobra a la tarifa por segundo de la tabla: si `duration` es `auto`, la insignia muestra esa tarifa por segundo; en caso contrario, el intervalo va desde la duración seleccionada hasta `min(duration + 15, 30)` segundos.
## WaveSpeed
@@ -1744,17 +1924,23 @@ Los endpoints de video cobran por el contenido moderado.
| :------------------------- | :--------- | :------- | :------------------------------------------------------------------ |
| grok-imagine-image-2.0 | 1K | generate | 12.66 (medium) / 8.44 (low) por imagen de salida |
| | 2K | generate | 16.88 (medium) / 12.66 (low) por imagen de salida |
-| grok-imagine-image-quality | 1K | generate | 10.55 por imagen de salida |
-| | 2K | generate | 14.77 por imagen de salida |
-| grok-imagine-image-pro | N/A | generate | 14.77 por imagen de salida |
+| grok-imagine-image-quality | 1K | generate | 8.44 por imagen de salida |
+| | 2K | generate | 12.66 por imagen de salida |
+| grok-imagine-image-pro | 1K | generate | 8.44 por imagen de salida |
+| | 2K | generate | 12.66 por imagen de salida |
| grok-imagine-image | N/A | generate | 4.22 por imagen de salida |
-| grok-imagine-image-2.0 | 1K | edit | 2.11 de entrada + 12.66 (medium) / 8.44 (low) por imagen de salida |
-| | 2K | edit | 2.11 de entrada + 16.88 (medium) / 12.66 (low) por imagen de salida |
-| grok-imagine-image-quality | 1K | edit | 2.11 de entrada + 10.55 por imagen de salida |
-| | 2K | edit | 2.11 de entrada + 14.77 por imagen de salida |
-| grok-imagine-image-pro | N/A | edit | 2.11 de entrada + 14.77 por imagen de salida |
+| grok-imagine-image-2.0 | 1K | edit | 2.11 a 6.33 de entrada + 12.66 (medium) / 8.44 (low) por imagen de salida |
+| | 2K | edit | 2.11 a 6.33 de entrada + 16.88 (medium) / 12.66 (low) por imagen de salida |
+| grok-imagine-image-quality | 1K | edit | 2.11 a 6.33 de entrada + 8.44 por imagen de salida |
+| | 2K | edit | 2.11 a 6.33 de entrada + 12.66 por imagen de salida |
+| grok-imagine-image-pro | 1K | edit | 2.11 de entrada + 8.44 por imagen de salida |
+| | 2K | edit | 2.11 de entrada + 12.66 por imagen de salida |
| grok-imagine-image | N/A | edit | 0.422 a 1.266 de entrada + 4.22 por imagen de salida |
+En los modelos de edición que no son pro, el cargo de entrada es de **2.11 créditos** por imagen de entrada, para **1 a 3** imágenes (**2.11–6.33**). `grok-imagine-image` usa **0.422 créditos** por imagen de entrada (**0.422–1.266**). Las filas de edición pro cobran exactamente una imagen de entrada a **2.11 créditos**.
+
+`grok-imagine-image-quality` y `grok-imagine-image-pro` están retirados por xAI. Los nodos de imagen aún los listan y los cobran a la tarifa de calidad baja de `grok-imagine-image-2.0`.
+
### Video
| modelo | resolución | Créditos |
@@ -1772,6 +1958,8 @@ Los endpoints de video cobran por el contenido moderado.
| | 720p | 21.12 / segundo de salida + 0.603 por imagen de referencia |
| grok-imagine-video (extend) | N/A | 6.03 a 45.26 de entrada + 15.09 × segundos de salida de la extensión / run |
+**Extend** cobra **15.09 créditos** por segundo de extensión, más un cargo de entrada de **6.03 créditos** si el vídeo de origen dura 2 segundos o **45.26 créditos** si dura 15 segundos (el origen debe durar entre 2 y 15 segundos). La insignia en vivo no puede ver esa duración, así que muestra esa banda de entrada **6.03–45.26** más el cargo por segundo de extensión.
+
## GPU en la nube
| gpu_type | Créditos |
diff --git a/ja/tutorials/partner-nodes/pricing.mdx b/ja/tutorials/partner-nodes/pricing.mdx
index e81540bbb2..4b54e10c05 100644
--- a/ja/tutorials/partner-nodes/pricing.mdx
+++ b/ja/tutorials/partner-nodes/pricing.mdx
@@ -3,54 +3,56 @@ title: "Partner Node 価格"
description: "Anthropic、OpenAI、Kling、Luma など全プロバイダーの Partner Node クレジット価格。モデル、トークン種別、生成種別ごとに分類。"
sidebarTitle: "料金"
mode: "wide"
-translationSourceHash: 2ef8ce31
+translationSourceHash: f68d51fd
translationFrom: tutorials/partner-nodes/pricing.mdx
translationBlockHashes:
"_intro": 1ec20604
- "Anthropic": 61f55fec
+ "Anthropic": 1757fac0
"Beeble": d1b19f61
- "BFL": 7ae11dc6
+ "BFL": 0a63f1b4
"Bria": 94080005
- "ByteDance": d2af4769
+ "Breeze": f38692c6
+ "ByteDance": fec72d04
"Clarity AI": 3d568b52
"ElevenLabs": cadb48cc
- "FishAudio": c957b7da
+ "FishAudio": 24ec3ea0
"Magnific": 4000e7e1
- "Google": b8be5b63
+ "Google": 958913c2
"HappyHorse": bceb584a
- "HeyGen": 7f08ab4e
+ "HeyGen": 248354a7
"Higgsfield": eed76dcb
"Hitpaw": 320c0537
- "Ideogram": da793624
+ "Ideogram": de0981df
"Krea": bcad7ba2
- "Kling": 32a38b1d
- "Lightricks": 3ba8d019
- "Luma": 9f809a47
+ "Kling": db91f830
+ "Lightricks": bbcb475e
+ "Luma": a2f2a8de
"Meshy": c057a289
"Meta": a253fe9a
"Mirelo": 2055c18a
- "Minimax": 07381eba
- "OpenAI": 17e24f34
- "OpenRouter": ef37d425
+ "Minimax": 3f9b84c5
+ "OpenAI": e51a94f8
+ "OpenRouter": 3de75ece
"Patina": d489c37e
- "Pixverse": bdb9b1c9
- "Quiver": 5538ce9c
+ "Pika": 66ba8620
+ "Pixverse": 86f21b3f
+ "Quiver": f84ffa99
"Qwen": 22dbe478
"Recraft": fc591c24
"Reve": 02eb39a9
"Rodin 3D": 2eb41834
"Runway": bd14d7ff
- "Sonilo": 48b9c581
+ "Sonilo": 66cc2806
"sync.so": 590402a9
"Tencent": cbe22aa6
- "Topaz": 5d08cf76
+ "Topaz": f557fb17
"Tripo": 842e861e
- "Pruna": f1b6a267
+ "Pruna": 08c63f88
"VEED": 1818bad3
"Vidu": be19d9ca
"Wan": 99df0ce3
"WaveSpeed": 349a7c9d
- "xAI": 28ff86de
+ "xAI": 6efd8650
"Cloud GPU": e228bb68
---
@@ -79,6 +81,7 @@ translationBlockHashes:
| :--------- | -----------------: | --------------------------: | --------------------------: | ---------------------: | :------------------ |
| Opus 5.5 | 1.207 | 1.509 | 2.414 | 0.121 | 6.035 |
| Opus 5、Opus 4.8 | 1.509 | 1.886 | 3.017 | 0.151 | 7.543 |
+| Fable 5.2 | 2.414 | 3.017 | 4.828 | 0.241 | 12.069 |
| Fable 5.1、Fable 5 | 3.017 | 3.772 | 6.035 | 0.302 | 15.087 |
| Sonnet 5.5、Sonnet 5 | 0.603 | 0.754 | 1.207 | 0.060 | 3.017 |
@@ -104,17 +107,19 @@ translationBlockHashes:
### 画像
-| ノード | クレジット |
-| :------------------------- | :---------------------------: |
-| FLUX 1.1 [pro] Ultra 画像 | 12.66 / 実行 |
-| FLUX.1 Kontext [pro] 画像 | 8.44 / 実行 |
-| FLUX.1 Kontext [max] 画像 | 16.88 / 実行 |
-| FLUX.1 展開 画像 | 10.55 / 実行 |
-| FLUX.1 Fill 画像 | 10.55 / 実行 |
-| FLUX 消去画像 | 6.33 / 実行 + 0.84 / 追加MP |
-| FLUX バーチャル試着 | 7.91 / 実行 + 1.06 / 追加MP |
-| FLUX.2 [pro] | 6.33 / 実行 + 3.17 / 追加MP |
-| FLUX.2 [max] | 14.77 / 実行 + 6.33 / 追加MP |
+| ノード | クレジット | クレジット(参照画像あり) |
+| :------------------------ | :--------------------------: | :----------------------------------: |
+| FLUX 1.1 [pro] Ultra 画像 | 12.66 / 実行 | — |
+| FLUX.1 Kontext [pro] 画像 | 8.44 / 実行 | — |
+| FLUX.1 Kontext [max] 画像 | 16.88 / 実行 | — |
+| FLUX.1 展開 画像 | 10.55 / 実行 | — |
+| FLUX.1 Fill 画像 | 10.55 / 実行 | — |
+| FLUX 消去画像 | 6.33 / 実行 + 0.84 / 追加MP | — |
+| FLUX バーチャル試着 | 7.91 / 実行 + 1.06 / 追加MP | — |
+| FLUX.2 [pro] | 6.33 / 実行 + 3.17 / 追加MP | 9.50 – 31.65 / 実行 + 3.17 / 追加MP |
+| FLUX.2 [max] | 14.77 / 実行 + 6.33 / 追加MP | 21.10 – 65.41 / 実行 + 6.33 / 追加MP |
+
+参照画像を接続すると 3 列目の料金になります。テキストのみの **FLUX.2 [pro]** は **6.33クレジット** に、出力の追加メガピクセル(`ceil(W × H / 1 MP) − 1`)ごとに **3.17クレジット** を加えた金額です。**FLUX.2 [max]** は **14.77クレジット** に、出力の追加メガピクセルごとに **6.33クレジット** を加えた金額です。参照画像がある場合は、入力メガピクセルごとに [pro] で **3.17クレジット**、[max] で **6.33クレジット** が加算されます(各参照画像の `ceil(W × H / 1 MP)` を合計、最大8枚)。3列目は出力が1 MP、参照入力が1〜8 MPのときの範囲(**9.50–31.65** / **21.10–65.41**)で、出力の追加メガピクセルはテキストのみの列と同じ単価で加算されます。[pro] の **3.17** は四捨五入した単価で、公表している **9.50–31.65** の範囲は参照入力 1〜8 MP のライブバッジ端点です。
### 画像: FLUX 3 Image
@@ -149,14 +154,14 @@ translationBlockHashes:
### 動画: FLUX 3 Max
-合計クレジット = **(クレジット / 秒)× `duration`**(`duration` は5〜20秒)。`reasoning_level` と `generate_audio` は価格に影響しません。ローンチプロモーション価格は **2026年11月7日(00:00 UTC)** まで適用され、その後は標準価格になります。
+合計クレジット = **(クレジット / 秒)× `duration`**(`duration` は5〜20秒)。`reasoning_level` と `generate_audio` は価格に影響しません。720p 12.82、1080p 21.88 クレジット / 秒のローンチ価格は 2026年11月7日(00:00 UTC)まで適用されました。以下の表は標準価格で、同日から適用されます。
-| ノード | 解像度 | プロモーション クレジット / 秒 | 標準 クレジット / 秒 |
-| :------------------------ | :---- | :---------------- | :----------- |
-| FLUX 3 Max Text to Video | 720p | 12.82 | 51.29 |
-| FLUX 3 Max Text to Video | 1080p | 21.88 | 87.50 |
-| FLUX 3 Max Image to Video | 720p | 12.82 | 51.29 |
-| FLUX 3 Max Image to Video | 1080p | 21.88 | 87.50 |
+| ノード | 解像度 | クレジット / 秒 |
+| :------------------------ | :---- | :-------- |
+| FLUX 3 Max Text to Video | 720p | 51.29 |
+| FLUX 3 Max Text to Video | 1080p | 87.50 |
+| FLUX 3 Max Image to Video | 720p | 51.29 |
+| FLUX 3 Max Image to Video | 1080p | 87.50 |
### 動画: Flux Video Edit
@@ -172,10 +177,10 @@ translationBlockHashes:
| ノード | モード | クレジット / MP / 秒 |
| :----------------- | :-------- | :----------------- |
-| Flux Video Upscale | creative | 15.09 |
-| Flux Video Upscale | precise | 10.56 |
+| Flux Video Upscale | creative | 14.39 |
+| Flux Video Upscale | precise | 10.07 |
-1080p出力(2.07MP)は creative で毎秒31.24クレジット、precise で21.86クレジットです。4K出力(8.29MP)はそれぞれ125.10と87.54になります。
+1080p出力(2.07MP)は creative で毎秒29.83クレジット、precise で20.88クレジットです。4K出力(8.29MP)はそれぞれ119.34と83.54になります。
## Bria
@@ -212,6 +217,18 @@ translationBlockHashes:
| Bria 動画背景除去(透過) | 10.55 / 秒 |
| Bria 動画消しゴム | 13.58 / 秒 |
+## Breeze
+
+Breeze の音声ノードはすべて **9.05 クレジット / 1K 課金文字** で課金され、CJK の文字は 2 文字として数えます。テキスト読み上げとボイスデザインは入力したテキストで課金されます。ボイスクローン、リミックス、ローカライズも同じ単価ですが、文字数は実行後にのみ分かります。リミックスとローカライズは生成されたプレビューテキスト、クローンはベンダーが報告する課金単位です。そのため、これら3つのノードのバッジはおよそ **0.66–2.66**、**0.72–1.52**、**0.76–2.29** クレジットの範囲を示します。
+
+| ノード | クレジット |
+| :------------------------ | :------------ |
+| Breeze テキスト読み上げ | 9.05 / 1K文字 |
+| Breeze ボイスデザイン | 9.05 / 1K文字 |
+| Breeze ボイスクローン | 9.05 / 1K文字 |
+| Breeze ボイスリミックス | 9.05 / 1K文字 |
+| Breeze ボイスローカライズ | 9.05 / 1K文字 |
+
## ByteDance
### LLM: ByteDance Seed
@@ -232,15 +249,17 @@ translationBlockHashes:
| モデル | クレジット / 画像 |
| :-------------------------------------------------- | :------------------- |
+| seedream 5.5(`seedream-5-5-261028`) | 6.03 (1K) / 10.56 (2K) |
| seedream 5.0 pro(`seedream-5-0-pro-260628`) | 9.50 (1K) / 18.99 (2K) |
| seedream 5.0 flash(`seedream-5-0-flash-260915`) | 5.43 |
| seedream 5.0 lite(`seedream-5-0-260128`) | 7.39 |
| seedream-4-5-251128 | 8.44 |
| seedream-4-0-250828 | 6.33 |
+| Seedream 5.5 レイヤー分離 | 6.03 (1K, 1.5K) / 10.56 (2K) |
| Seedream 5.0 Pro レイヤー分離 | 6.79 (1K, 1.5K) / 13.58 (2K) |
| Seedream 5.0 Flash レイヤー分離 | 5.43 |
-Seedream 5.0 Proは解像度ベースの課金方式を採用しています:**1K**(≤2.61MP、**9.50クレジット**)および**2K**(>2.61MP、**18.99クレジット**)。Seedream 5.0 Flash は解像度にかかわらず 1 枚あたり **5.43 クレジット** の定額です。各リクエストの最初の画像は無料です。最初の画像を超える追加画像(出力画像と入力/参照画像の両方を含む)には、**0.63クレジット**が加算されます。
+Seedream 5.0 Proは解像度ベースの課金方式を採用しています:**1K**(≤2.61MP、**9.50クレジット**)および**2K**(>2.61MP、**18.99クレジット**)。Seedream 5.0 Flash は解像度にかかわらず 1 枚あたり **5.43 クレジット** の定額です。各リクエストの最初の画像は無料です。最初の画像を超える追加画像(出力画像と入力/参照画像の両方を含む)には、**0.63クレジット**が加算されます。**Seedream 5.5** は約2.6MPまでが1枚 **6.03クレジット**、それを超えると **10.56クレジット**で、参照画像を1枚つなぐごとに **2.41クレジット**が加算されます。 **Seedream 5.5 レイヤー分離** は 1K/1.5K で **6.03クレジット**、2K で **10.56クレジット**、実行あたりの画像数で課金されます。
### 画像: ByteDance Image(非推奨)
@@ -254,24 +273,45 @@ Seedream 5.0 Proは解像度ベースの課金方式を採用しています:*
### 動画: Seedance 2.5
-トークン基準の課金。推定トークン消費量 = **(入力動画時間 + 出力動画時間)× 出力動画幅 × 出力動画高さ × 出力動画フレームレート ÷ 1024**。テキスト、画像、音声のみの入力の場合、入力動画時間は0です。実際の使用量はAPIを基準とします。
+トークン基準の課金。ComfyUI のライブ価格バッジはトークンを **floor((出力幅 × 出力高さ ÷ 1024)×(24 × 課金時間 + 1))** で見積もり、下表の「クレジット / 1Kトークン」料金を掛けます(`+ 1` は 1 フレーム分の加算です)。課金時間は出力時間であり、参照動画が接続されている場合は入力動画時間を足します。テキスト、画像、音声のみの入力の場合、入力動画時間は 0 です。実際の使用量は API を基準とします。
+
+| モデル | 解像度 | 動画入力 | クレジット / 1Kトークン |
+| :----------------- | :--------- | :------- | ----------------------: |
+| Seedance 2.5 | 480p, 720p | なし | 3.229 |
+| Seedance 2.5 | 480p, 720p | あり | 1.931 |
+| Seedance 2.5 | 1080p | なし | 3.530 |
+| Seedance 2.5 | 1080p | あり | 2.112 |
+| Seedance 2.5 | 4k | なし | 1.418 |
+| Seedance 2.5 | 4k | あり | 0.875 |
+| Seedance 2.5 Flash | 480p, 720p | なし | 2.112 |
+| Seedance 2.5 Flash | 480p, 720p | あり | 1.297 |
+
+**動画入力**は、参照動画が接続されていることを意味します。テキスト、画像、音声の参照は動画入力なしとしてカウントされます。ライブバッジは、上記の計算式以外に別途の最低トークン課金を適用しません。
+
+**Seedance 2.5 Flash** は 480p と 720p のみ、最長 30 秒で、どちらも同じ料金です。動画入力なしの 5 秒 720p は約 230 クレジットです。
+
+同じ長さでは 4k の 1 フレームあたりのトークン数が 1080p の約 4 倍になるため、1K トークンあたりの単価は低くなりますが、合計料金は高くなります。
+
+参照動画は2秒以上、入力の合計は最大30秒です。上のトークン計算式による、16:9出力(24fps)の推定料金:
-| モデル | 解像度 | 動画入力 | クレジット / 1Kトークン |
-| :----------- | :--------- | :------- | ----------------------: |
-| Seedance 2.5 | 480p, 720p | なし | 3.229 |
-| Seedance 2.5 | 480p, 720p | あり | 1.931 |
-| Seedance 2.5 | 1080p | なし | 3.530 |
-| Seedance 2.5 | 1080p | あり | 2.112 |
+| 解像度 | 5秒出力、動画入力なし | 5秒出力 + 2秒の動画入力 | 5秒出力 + 30秒の動画入力 |
+| :--------- | -------------------: | ----------------------: | -----------------------: |
+| 480p | 156.38 | 130.64 | 650.12 |
+| 720p | 351.58 | 293.72 | 1461.63 |
+| 1080p | 865.00 | 722.82 | 3596.98 |
+| 4k | 1389.91 | 1197.81 | 5960.70 |
-**動画入力**は、参照動画が接続されていることを意味します。テキスト、画像、音声の参照は動画入力なしとしてカウントされます。入力に動画が含まれる場合、最低トークン消費量の制限が適用されます:推定消費量が最低値を下回る場合、最低消費量で課金されます。
+参照動画を接続すると 1K トークンあたりの単価が下がるため、5秒出力に2秒の参照を加えた料金は、動画入力なしの5秒より安くなることがあります。ライブバッジは入力動画の長さを参照できないため、参照動画ありの5秒実行では、「出力5秒のみ(動画入力ありの単価)」から「出力5秒 + 入力30秒」までの範囲を表示します。
-参照動画は2秒以上、入力の合計は最大30秒です。16:9出力(24fps)の推定料金:
+**Seedance 2.5 Draft** モデルで実行すると、上の最初の表の 480p トークンレートで 480p プレビューを生成し、`draft_task_id` を返します。**ByteDance Seedance 2.5 Draft to Final Video** はそのドラフトを 1080p または 4k の最終動画としてレンダリングします。最終動画はドラフトのプロンプト、参照、長さ、音声設定を引き継ぎます。このノードのライブバッジは、同じ計算式による時間に依存しないクレジット範囲を表示します(下限は参照動画なしの 4 秒 16:9 最終動画、上限は 30 秒の最終動画に 30 秒の参照動画を加えた場合です)。上限は 1080p では 21:9、4k では 4:3 または 3:4 です:
+
+| ノード | 解像度 | クレジット |
+| :------------------------------------------ | :----- | ----------------: |
+| ByteDance Seedance 2.5 Draft to Final Video | 1080p | 693.43 – 6202.64 |
+| ByteDance Seedance 2.5 Draft to Final Video | 4k | 1114.23 – 10214.09 |
+
+参照動画なしの 4 秒の 1080p・16:9 最終動画は 693.43 クレジット、30 秒の参照動画を伴う 30 秒の 1080p・16:9 最終動画は 6163.19 クレジットです。どちらもバッジの追加 1 フレームを含みます。アスペクト比が変わると料金も少し変わります。1080p の上限 6202.64 は 21:9、4k の上限 10214.09 は 4:3 または 3:4 の場合です。最終的な課金は API の実際の使用量で決まります。
-| 解像度 | 5秒出力、動画入力なし | 5秒出力 + 2〜30秒の動画入力 |
-| :--------- | -------------------: | -------------------------: |
-| 480p | 155.1 | 166.9 – 649.3 |
-| 720p | 348.8 | 375.4 – 1459.8 |
-| 1080p | 857.8 | 923.9 – 3592.7 |
### 動画: Seedance 2.0
480pと720pは同一レートです。FastおよびMiniは1080pまたは4Kをサポートしていません。
@@ -311,7 +351,7 @@ Seedream 5.0 Proは解像度ベースの課金方式を採用しています:*
| seedance-1-0-pro-fast-251015 | — | 211 |
- **非推奨**: `seedance-1-5-pro-251215` は BytePlus が同モデルを停止するため 2026年11月11日 に提供終了となり、ComfyUI も同日に ByteDance 1.x ノードから削除します。そのため本ページからもこのモデルの価格行を削除しました。代わりに `seedance-1-0-pro-250528`、`seedance-1-0-pro-fast-251015`、または Seedance 2.x モデルを使用してください。
+ **提供終了**: `seedance-1-5-pro-251215` は BytePlus が 2026年11月11日 に停止したため、本ページからこのモデルの価格行を削除しました。代わりに `seedance-1-0-pro-250528`、`seedance-1-0-pro-fast-251015`、または Seedance 2.x モデルを使用してください。
**提供終了済み**: `seedance-1-0-lite-t2v-250428` と `seedance-1-0-lite-i2v-250428`、および **ByteDance 参照画像 to 動画** ノードは提供終了となり、ComfyUI から削除されました。以下の価格は参考としてのみ残されています。
@@ -403,7 +443,7 @@ Seedream 5.0 Proは解像度ベースの課金方式を採用しています:*
## FishAudio
-**期間限定オファー**: 2026 年 10 月 13 日から 11 月 12 日まで、Drama 3 は約 **3.02** クレジット / 1K バイトで課金されます。表の `drama-3` 行は標準価格で、2026 年 11 月 13 日から適用されます。
+Drama 3 は 2026 年 10 月 13 日から 11 月 12 日まで、導入レートの約 **3.02** クレジット / 1K バイトで課金されました。表の `drama-3` 行は標準価格で、2026 年 11 月 13 日から適用されます。
| ノード | モデル | クレジット |
@@ -446,6 +486,9 @@ Seedream 5.0 Proは解像度ベースの課金方式を採用しています:*
| モデル | 入力クレジット / 1K | 出力クレジット / 1K |
| :------------------------------------------------------ | -----------------: | :------------------ |
+| Gemini 4 Hydrogen(`gemini-4-hydrogen`) | 0.4526 | 2.263 |
+| Gemini 4 Argon(`gemini-4-argon`) | 1.20692 | 6.0346 |
+| Gemini 4 Carbon(`gemini-4-carbon`) | 1.20692 | 6.0346 |
| Gemini 3.8 Flash(`gemini-3.8-flash`) | 0.4526 | 2.263 |
| Gemini 3.7 Flash(`gemini-3.7-flash`) | 0.4526 | 2.263 |
| Gemini 3.5 Flash(`gemini-3.5-flash`) | 0.3165 | 1.899 |
@@ -477,6 +520,10 @@ Seedream 5.0 Proは解像度ベースの課金方式を採用しています:*
| | 720p | 30.57 |
| | 1080p | 45.87 |
| | 4k | 91.76 |
+| Gemini Omni Flash 1.2(`gemini-omni-1.2-flash`) | 360p | 6.98 |
+| | 720p | 20.97 |
+| | 1080p | 31.46 |
+| | 4k | 62.92 |
### 動画: Google Veo
@@ -532,23 +579,33 @@ Seedream 5.0 Proは解像度ベースの課金方式を採用しています:*
| HeyGen Video Translate | スピードモード | 4.07 / 秒 |
| HeyGen Text to Speech | N/A | ~0.20 / 秒 |
| HeyGen Video 1.0 Reference to Video | 480p、参照素材なし | 3.02 / 秒 |
-| HeyGen Video 1.0 Reference to Video | 768p、参照素材なし | 4.53 / 秒 |
-| HeyGen Video 1.0 Reference to Video | 2K、参照素材なし | 13.58 / 秒 |
-| HeyGen Video 1.0 Reference to Video | 480p、参照素材あり | 6.03 / 秒 |
-| HeyGen Video 1.0 Reference to Video | 768p、参照素材あり | 9.05 / 秒 |
-| HeyGen Video 1.0 Reference to Video | 2K、参照素材あり | 27.16 / 秒 |
-| HeyGen Video 1.0 Image to Video | 480p | 3.02 / 秒 |
-| HeyGen Video 1.0 Image to Video | 768p | 4.53 / 秒 |
-| HeyGen Video 1.0 Image to Video | 2K | 13.58 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 768p、参照素材なし | 5.43 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 1080p、参照素材なし | 12.07 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 2K、参照素材なし | 15.09 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 480p、参照素材あり | 10.56 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 768p、参照素材あり | 15.09 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 1080p または 2K、参照素材あり | 27.16 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 3枚目以降の参照画像 | 6.03 / 枚 |
+| HeyGen Video 1.0 Image to Video | 480p | 3.02 / 秒 |
+| HeyGen Video 1.0 Image to Video | 768p | 5.43 / 秒 |
+| HeyGen Video 1.0 Image to Video | 1080p | 12.07 / 秒 |
+| HeyGen Video 1.0 Image to Video | 2K | 15.09 / 秒 |
+| HeyGen Video 1.0 Image to Video | 15 秒超、480p | 13.58 / 秒 |
+| HeyGen Video 1.0 Image to Video | 15 秒超、768p | 22.63 / 秒 |
+| HeyGen Video 1.0 Image to Video | 15 秒超、1080p または 2K | 31.68 / 秒 |
| HeyGen Voice 1.0 Instant Clone | N/A | 無料 |
| HeyGen Voice 1.0 Text to Speech | `heygen-voice-1` | 9.05 / 1K文字 |
| HeyGen Voice 1.0 Text to Speech | `heygen-voice-1-turbo` | 2.26 / 1K文字 |
-**テキスト読み上げの料金**: 2026 年 10 月 31 日までは、1K 文字あたり約 **4.53** クレジット(`heygen-voice-1`)、**1.21** クレジット(`heygen-voice-1-turbo`)の提供開始料金が適用されました。上表は標準価格で、2026 年 11 月 1 日から適用されます。
+**テキスト読み上げの料金**: 2026 年 11 月 1 日(00:00 UTC)までは、1K 文字あたり約 **4.53** クレジット(`heygen-voice-1`)、**1.21** クレジット(`heygen-voice-1-turbo`)の提供開始料金が適用されました。上表は標準価格で、その時点から適用されます。
アバタービデオは、選択したアバタールックのタイプ(Digital Twin / Photo Avatar / Studio Avatar)のレートで秒単位で課金されます。**オート**はルックに応じてエンジンを自動選択するため、レートは上記の全範囲に及びます。カスタムアバターID はそのタイプに応じて課金されます。
-HeyGen Video 1.0 のクリップは 5〜15 秒で、選択した解像度のレートで課金されます。Reference to Video は参照画像または参照動画を接続するまで参照素材なしのレートで課金され、参照動画 1 本ごとに課金対象の長さが 5 秒増えます。**2K** 出力には `16:9` または `9:16` のアスペクト比が必要です。`auto` は参照メディアを接続していない場合にのみ `2K` で使用できます。
+**期間限定**: 50%オフのプロモーション料金は 2026 年 12 月 1 日(00:00 UTC)まで適用されます。上表の Video 1.0 の行は現行のプロモーション料金です。長尺クリップのレートはすでに標準価格です。2026 年 12 月 1 日(00:00 UTC)以降は標準価格が適用されます。参照素材なしおよび Image to Video(15秒以下)は毎秒 **6.03 / 10.86 / 24.14 / 30.17** クレジット、参照素材ありは毎秒 **21.12 / 30.17 / 54.31** クレジット、3枚目以降の参照画像は1枚あたり **12.07クレジット** です。長尺クリップのレートは変わりません。
+
+HeyGen Video 1.0 のクリップは 5〜30 秒で、秒単位で課金されます。15 秒を超えるクリップは、両ノードとも参照素材を接続していない場合に長尺の行のレートで課金されます。参照素材を接続している場合、15秒超は利用できません。画像または動画を1つでも接続すると、最初の2枚に1枚あたりの追加料金がなくても、**Reference to Video** は参照素材ありのレートに切り替わります。最初の2枚の参照画像は無料で、3枚目以降は1枚あたり **6.03クレジット** が加算されます(この加算は 2026 年 12 月 1 日(00:00 UTC)以降 **12.07クレジット** になります)。参照映像は合計15秒まで、参照素材ありの秒単価で課金されます。
+
+`auto_duration` は `duration` の代わりにプロンプト内のセリフの長さを使用します(8〜30 秒、参照素材を接続した場合は 8〜15 秒に参照映像の長さが加算されます)。**1080p** と **2K** 出力には `16:9` または `9:16` のアスペクト比が必要で、`auto` は参照メディアがない場合にのみこれらの解像度で使用できます。
## Higgsfield
@@ -659,15 +716,15 @@ HeyGen Video 1.0 のクリップは 5〜15 秒で、選択した解像度のレ
## Ideogram
-| ノード | ターボ | キャラクター参照 | レンダリング速度 | クレジット / 画像 |
-| :---------- | :---- | :------------------ | :-------------- | :-------------- |
-| Ideogram V1 | false | — | — | 18.1 |
-| Ideogram V1 | true | — | — | 6.03 |
-| Ideogram V2 | false | — | — | 24.14 |
-| Ideogram V2 | true | — | — | 15.09 |
-| Ideogram V4 | — | — | TURBO | 9.05 |
-| Ideogram V4 | — | — | DEFAULT | 18.1 |
-| Ideogram V4 | — | — | QUALITY | 30.17 |
+
+ **非推奨**: ComfyUI から **Ideogram V1** と **Ideogram V2** ノードが削除されたため、本ページではこれらの価格を掲載していません。代わりに **Ideogram V4** を使用してください。
+
+
+| ノード | レンダリング速度 | クレジット / 画像 |
+| :---------- | :--------------- | :---------------: |
+| Ideogram V4 | TURBO | 9.05 |
+| Ideogram V4 | DEFAULT | 18.1 |
+| Ideogram V4 | QUALITY | 30.17 |
| ノード | 品質 | 解像度 | クレジット / 画像 |
| :--------------- | :------ | :----- | :---------------- |
@@ -726,13 +783,41 @@ HeyGen Video 1.0 のクリップは 5〜15 秒で、選択した解像度のレ
| Krea 3 | 2K | high | 24.14 | 6.03 - 21.12 |
| Krea 3 | 2K | ultra | 78.45 | 6.03 - 21.12 |
-2K 出力では、参照画像ごとに参照画像自体の画素サイズに応じて **6.03〜21.12 クレジット** が加算されます。1K 出力では参照画像が約 1 MP にダウンサンプリングされるため、加算は **6.03 クレジット** の定額です。`ultra` は `resolution` の値にかかわらず常に 2K で出力され、参照画像が 1 枚以上必要です。
+2K 出力では、参照画像ごとに **6.03〜21.12クレジット** が加算されます(下限 **6.03**、上限 **21.12**。ノードは参照画像の画素サイズに応じてこの2つの値の間で課金します)。1K 出力では参照画像が約 1 MP にダウンサンプリングされるため、加算は **6.03 クレジット** の定額です。`ultra` は `resolution` の値にかかわらず常に 2K で出力され、参照画像が 1 枚以上必要です。
## Kling
### 動画: クレジット / 秒
-合計クレジット = **(クレジット / 秒)× `duration`**。ただし **Kling 3.0 Omni Edit Video** と **Kling Motion Control** を除く(出力の秒単位で課金。Editでは時間は入力動画に従う)。
+合計クレジット = **(クレジット / 秒)× `duration`**。ただし **Kling 3.0 Omni Edit Video**、**Kling 4.0 Edit Video** と **Kling Motion Control** を除く(出力の秒単位で課金。Editでは時間は入力動画に従う)。
+
+#### kling-4.0
+
+`kling-4.0` は 3〜30 秒、720p または 1080p の動画を生成します。`kling-4.0-flash` は 3〜20 秒、720p です。
+
+| ノード | モデル | 解像度 | クレジット / 秒 | 参照動画ありのクレジット / 秒 |
+| :--------------------------- | :-------------- | :----- | :-------------- | :---------------------------- |
+| Kling 4.0 Text to Video | kling-4.0 | 720p | 47.37 | — |
+| Kling 4.0 Text to Video | kling-4.0 | 1080p | 70.91 | — |
+| Kling 4.0 Image to Video | kling-4.0 | 720p | 47.37 | — |
+| Kling 4.0 Image to Video | kling-4.0 | 1080p | 70.91 | — |
+| Kling 4.0 Reference to Video | kling-4.0 | 720p | 47.37 | 28.36 |
+| Kling 4.0 Reference to Video | kling-4.0 | 1080p | 70.91 | 42.54 |
+| Kling 4.0 Text to Video | kling-4.0-flash | 720p | 28.36 | — |
+| Kling 4.0 Image to Video | kling-4.0-flash | 720p | 28.36 | — |
+| Kling 4.0 Reference to Video | kling-4.0-flash | 720p | 28.36 | 19.01 |
+
+**Kling 4.0 Reference to Video** で参照動画を接続すると、実行全体が低いレートに切り替わります。出力の各秒と参照動画の各秒の両方がそのレートで課金されます。各クリップは 3〜30 秒(`kling-4.0-flash` は 3〜20 秒)で、合計も同じ上限内に収める必要があります。5 秒の 720p 出力に 30 秒の参照動画を付けると 992.7 クレジットです。
+
+**Kling 4.0 Edit Video** は常に低い(参照動画の)レートで課金されます。**クレジット = 単価 ×(入力動画の秒数 + 参照動画の秒数)**。出力は入力と同じ長さになるため、参照動画がない場合は入力1秒あたり **単価の2倍** です。参照動画がある場合も実際の料金はこの式に従います。ライブバッジは動画の長さを参照できないため、入力1秒あたり **1倍〜2倍** の範囲を表示します。参照画像に1枚あたりの追加料金はありません。
+
+| モデル | 解像度 | 単価(クレジット / 秒) | 参照動画なし | 参照動画あり |
+| :-------------- | :----- | ---------------------: | :----------- | :----------- |
+| kling-4.0 | 720p | 28.36 | 56.73 / 入力1秒(2 × 単価) | 28.36 ×(入力 + 参照秒数) |
+| kling-4.0 | 1080p | 42.54 | 85.09 / 入力1秒(2 × 単価) | 42.54 ×(入力 + 参照秒数) |
+| kling-4.0-flash | 720p | 19.01 | 38.02 / 入力1秒(2 × 単価) | 19.01 ×(入力 + 参照秒数) |
+
+参照動画なしの 10 秒 1080p 編集は 850.9 クレジットです。
#### kling-video-o1
@@ -788,10 +873,6 @@ HeyGen Video 1.0 のクリップは 5〜15 秒で、選択した解像度のレ
| motion control | — | std | 14.77 |
| motion control | — | pro | 23.63 |
-
- **サービス終了(EOL)**:Kling V1.5、V1.6、V2.1、V2.1 Master、および Kolors バーチャル試着は **2026年9月15日** に廃止予定です。以下の料金表は削除まで参考用として残しています。
-
-
#### kling-v2-5-turbo
**Kling Text to Video**、**Kling Image (First Frame) to Video**、**Kling Start-End Frame to Video** は `kling-v2-5-turbo` の **pro** モードのみを使用し、固定の5秒または10秒の実行単位で課金されます。
@@ -839,6 +920,10 @@ HeyGen Video 1.0 のクリップは 5〜15 秒で、選択した解像度のレ
| :------------------- | :-------------- |
| Kling Virtual Try-On | 33.79 |
+
+ **廃止済み**: Kling は **2026年9月15日** に V1.5、V1.6、V2.1、V2.1 Master、および以前の Kolors バーチャル試着サービスを廃止しました。Kling ノードはこれらのモデルを提供しなくなり、このページからも料金行を削除しています。上記の **Kling Virtual Try-On** の行は別の現行ノードで、1 回ごとの課金です。
+
+
## Lightricks
### 動画: クレジット / 秒
@@ -868,15 +953,35 @@ HeyGen Video 1.0 のクリップは 5〜15 秒で、選択した解像度のレ
解像度は `1280x720` / `720x1280`(720p)、`1920x1080` / `1080x1920`(1080p)、`2560x1440` / `1440x2560`(1440p)、`3840x2160` / `2160x3840`(4K)に対応します。**LTX-2.5(Fast)** で10秒を超える時間は720pまたは1080p解像度と24/25 FPSが必要です。**LTX-2(Fast)** で10秒を超える時間は1080pおよび25 FPSでのみ利用可能です。
+### ポストプロセス: クレジット / 秒
+
+これらのノードは入力動画を処理し、その秒数に応じて課金されます。
+
+| ノード | 解像度 | クレジット / 秒 |
+| :------------------- | ------ | --------------- |
+| LTX Video AlphaGen | 1080p(短辺 ≤1080。1080未満はこの料金) | 30.17 |
+| LTX Video AlphaGen | 1440p(短辺 ≤1440) | 75.43 |
+| LTX Video AlphaGen | 2160p(短辺 ≤2160) | 301.73 |
+| LTX Video SDR to HDR | 1080p(≤ 1920×1080) | 120.69 |
+| LTX Video SDR to HDR | 1440p(≤ 2560×1440) | 241.39 |
+| LTX Video SDR to HDR | 4K(≤ 3840×2160) | 482.77 |
+| LTX Video Reframe | 720p | 30.17 |
+| LTX Video Reframe | 1080p | 60.35 |
+| LTX Video Upscale | 1080p | 24.14 |
+| LTX Video Upscale | 4k | 90.52 |
+| LTX Video Upscale | 8k | 362.08 |
+
+**AlphaGen** は入力動画の秒数で課金され、料金帯は入力の短辺(幅と高さのうち小さい方)で決まります。短辺が1080ピクセル未満の入力は1080p料金になります。**SDR to HDR** も入力動画の秒数で課金され、入力の画素数を収めることができる最小の解像度帯にスナップします。ライブの ComfyUI バッジは入力解像度を参照できないため、範囲のみを表示します(AlphaGen は **30.17–301.73** クレジット/秒、SDR to HDR は **120.69–482.77** クレジット/秒)。実際の課金は表の 3 段階の料金です。**AlphaGen** は最大10秒、**Reframe** は最大60秒の動画を受け付け、**SDR to HDR** と **Upscale** は解像度ごとにフレーム数を制限します(SDR to HDR は1080pで273フレーム、Upscale は2,401フレーム)。
+
## Luma
### 動画
- **非推奨**: 旧版の **Luma Text to Video** および **Luma Image to Video** ノードは非推奨です。Luma はこれらのノードが使用するエンドポイントを 2026年10月24日 に削除したため、この 2 つのノードは映像を生成できなくなりました。代わりに **Luma Ray 3.2** ノードを使用してください。以下の ray-2、ray-flash-2、ray-1-6 の行は参考用としてのみ残されています。**ray-1-6** は 2026年9月以降提供されていません。
+ **廃止済み**: 旧版の Luma Text to Video、Image to Video、画像生成ノードは削除され、ray-2、ray-flash-2、ray-1-6、photon-1、photon-flash-1 モデルも提供されなくなりました。以下の表には ComfyUI が現在提供しているノードのみを記載しています。
-各 `duration` 選択肢について、フラットな **/実行** 価格がノードバッジと一致します。**Edit** および **reframe** は **(クレジット / 秒)× 動画長** で課金されます。**ray-1-6** は常に **105.50 / 実行** です。
+各 `duration` 選択肢について、フラットな **/実行** 価格がノードバッジと一致します。**Edit** および **reframe** は **(クレジット / 秒)× 動画長** で課金されます。
| モデル | タスク | 解像度 | クレジット |
| :---------- | :--------- | :--------- | :------------------------------------------------ |
@@ -892,24 +997,13 @@ HeyGen Video 1.0 のクリップは 5〜15 秒で、選択した解像度のレ
| ray-3.2 | reframe | 540p | 12.66 / 秒 |
| ray-3.2 | reframe | 720p | 25.32 / 秒 |
| ray-3.2 | reframe | 1080p | 75.96 / 秒 |
-| ray-2 | generation | 540p | 120.27 / 実行 (5s), 217.33 / 実行 (9s) |
-| ray-2 | generation | 720p | 215.22 / 実行 (5s), 386.13 / 実行 (9s) |
-| ray-2 | generation | 1080p | 478.97 / 実行 (5s), 865.10 / 実行 (9s) |
-| ray-2 | generation | 4K | 1922.21 / 実行 (5s), 3460.40 / 実行 (9s) |
-| ray-flash-2 | generation | 540p | 42.20 / 実行 (5s), 75.96 / 実行 (9s) |
-| ray-flash-2 | generation | 720p | 71.74 / 実行 (5s), 128.71 / 実行 (9s) |
-| ray-flash-2 | generation | 1080p | 166.69 / 実行 (5s), 299.62 / 実行 (9s) |
-| ray-flash-2 | generation | 4K | 660.43 / 実行 (5s), 1192.15 / 実行 (9s) |
-| ray-1-6 | generation | — | 105.50 / 実行 |
-
-**Luma Ray 3.2** Image to Video および Extend Video は常に **5秒** の価格で課金されます。レガシー Image to Video は **ray-2** / **ray-flash-2** 行の **5秒** または **9秒** を使用します。
+
+**Luma Ray 3.2** Image to Video および Extend Video は常に **5秒** の価格で課金されます。
### 画像
| モデル | タスク | クレジット |
| :------------- | :--------- | :----------------------------------- |
-| photon-flash-1 | — | 0.57 / 実行 |
-| photon-1 | — | 2.19 / 実行 |
| uni-1 | generation | 8.52 / 実行 + 0.63 / 参照画像(≤ 9) |
| uni-1 | edit | 9.16 / 実行 + 0.63 / 参照画像(≤ 8) |
| uni-1-max | generation | 21.10 / 実行 + 0.63 / 参照画像(≤ 9) |
@@ -975,7 +1069,7 @@ Ultra モード(Meshy 7.1、Meshy 7、`latest` モデルで利用可能)は
| MiniMax H3 最初と最後のフレームからビデオ | MiniMax H3 | 2K | 39.22 / 秒 |
| MiniMax H3 リファレンスからビデオ | MiniMax H3 | 768P | 27.16 / 秒、5枚を超える画像は1枚につき+12.07 |
| MiniMax H3 リファレンスからビデオ | MiniMax H3 | 2K | 39.22 / 秒、5枚を超える画像は1枚につき+12.07 |
-| MiniMax H3 コンテキストIR | MiniMax H3 | — | 4.22 / 実行(テキストのみ)、10.55 / 実行(最初と最後のフレーム)、12.66–23.21 / 実行(参照メディア) |
+| MiniMax H3 コンテキストIR | MiniMax H3 | — | 271.56 / 入力トークン100万 + 1,086.23 / 出力トークン100万 |
| MiniMax H3 2K再生成 | MiniMax H3 | — | 15.09 / 秒 |
| MiniMax H3 テキストからビデオ | MiniMax H3 Max | 480P | 15.09 / 秒 |
| MiniMax H3 テキストからビデオ | MiniMax H3 Max | 768P | 24.14 / 秒 |
@@ -1008,7 +1102,7 @@ Ultra モード(Meshy 7.1、Meshy 7、`latest` モデルで利用可能)は
**MiniMax Hailuo 02 Video** は1080Pでは **6秒** のみサポートします。
-**MiniMax H3** は秒単位で課金されます(4〜15秒)。**MiniMax H3 リファレンスからビデオ** は、5枚を超える参照画像1枚につき12.07クレジット、参照ビデオは27.16クレジット/秒(768P)または39.22クレジット/秒(2K)で課金されます(1本あたり2秒から、合計最大15秒、概算)。**MiniMax H3 コンテキストIR** の料金は添付メディア(テキストのみ、最初と最後のフレーム、参照メディア)によって異なります(概算)。
+**MiniMax H3** は秒単位で課金されます(4〜15秒)。**MiniMax H3 リファレンスからビデオ** は、5枚を超える参照画像1枚につき12.07クレジット、参照ビデオは27.16クレジット/秒(768P)または39.22クレジット/秒(2K)で課金されます(1本あたり2秒から、合計最大15秒、概算)。**MiniMax H3 コンテキストIR** は、**入力トークン100万あたり 271.56クレジット**、**出力トークン100万あたり 1,086.23クレジット** で課金されます。ノードバッジはグラフ作成時にトークン数を把握できないため、おおよその範囲として **4.22**(テキストのみ)、**10.55**(最初と最後のフレーム)、**12.66–23.21**(参照メディア)を表示します。
**MiniMax H3 Max** は上表の秒単価で課金されます。**テキストからビデオ**、**最初と最後のフレームからビデオ**、**リファレンスからビデオ** の出力は **5〜15秒** です。**リップシンク** は 5〜約899秒の音声に対応し、出力が15秒を超える場合は下記の **1.2×** 倍率が適用されます。
@@ -1034,6 +1128,19 @@ Ultra モード(Meshy 7.1、Meshy 7、`latest` モデルで利用可能)は
**MiniMax H3 Max 人物置換**、**動画延長**、**リフレーム**、**動画挿入**、**リップシンク**、**カメラコントロール** は動画の秒数に応じて課金されます。動画延長は Turbo モデルも選べ、料金は Max の半分です。**リフレーム**は表のレートで秒単位に課金されます。長いソースは重なるウィンドウを使うため、課金秒数が最終出力の長さを超えることがあります。**リップシンク**は表のレートで秒単位に課金されます。出力が**15秒を超える**場合、**1.2×** の倍率が適用されます(例: 768P → 28.97 / 秒)。動画挿入は無料の4,096トークン枠を超えた参照トークンに1Kトークンあたり6.03クレジットが加算され、リファレンスからビデオと同じ枠を共有します(基本240トークン、画像1枚2,048トークン、参照動画は768P出力で225,000トークン、480Pで87,000トークン)。
+### 画像: MiniMax H3 Image
+
+生成された画像1枚あたり。出力サイズに加え、プロンプトと参照画像がトークンとして課金されます。
+
+| ノード | モデル | サイズ | クレジット / 画像 |
+| :--------------- | -------------------------- | ------ | ----------------- |
+| MiniMax H3 Image | MiniMax-H3-Image | 2K | 5.32 – 8.17 |
+| MiniMax H3 Image | MiniMax-H3-Image | 4K | 19.41 – 22.26 |
+| MiniMax H3 Image | MiniMax-H3-Image-Highspeed | 2K | 10.02 – 12.86 |
+| MiniMax H3 Image | MiniMax-H3-Image-Highspeed | 4K | 38.20 – 41.04 |
+
+上表はテキストのみで `enhance_prompt` がオンの場合の料金です。完全な計算式は次のとおりです。出力クレジット = `max(ピクセル数, 4,194,304) / 1,024 × 0.00114657`(Highspeed では `0.00229315`)。2K(4 MP)は **4.70クレジット**、4K は約 **18.79クレジット**、Highspeed では **9.39** / **37.57** です。プロンプトトークンは1トークンあたり **0.000694クレジット** で、参照画像なし・拡張オフは **0〜300**、参照画像なし・拡張オンは **900〜5,000**、参照画像が `n` 枚のときは **2,000 + 300 × (n − 1)** から **6,500 + 800 × (n − 1)** です。各参照画像はさらに **2,050〜8,194** の画像トークンを、1トークンあたり **0.000453クレジット**(Highspeed では **0.000905**)で加算し、約 **0.93〜3.71クレジット**(Highspeed では **1.86〜7.42クレジット**)になります。拡張オフの2Kは1枚あたり約 **4.70クレジット** です。カスタムサイズはその面積で計算され、下限は4メガピクセルです。
+
## OpenAI
DALL·E と Sora は固定の画像あたりまたは秒あたりのレートを使用します。**OpenAI GPT Image 2** および **OpenAI ChatGPT** は API トークン使用量に基づいて課金されます。
@@ -1084,14 +1191,55 @@ DALL·E と Sora は固定の画像あたりまたは秒あたりのレートを
出力トークン数はサイズと品質に応じて変動します(例:1024x1024 の `low` は約196出力トークンで、`gpt-image-1.5` では約1.32クレジット)。1024x1024 の参照画像 1 枚あたり約 2.47 クレジットの画像入力が追加され、プロンプトテキストはテキスト入力レートで課金されます。生成枚数(`n`)に応じて合計が比例します。
+**GPT Image 2.6** モデル(`gpt-image-2.6-flare`、`gpt-image-2.6-sunburst`)は同じレート表を使用します。以下は 1024x1024 の画像 1 枚あたりのクレジットです。
+
+| 品質 | クレジット / 枚 |
+| :----- | :-------------- |
+| low | 7.45 |
+| medium | 13.12 |
+| high | 15.89 |
+| xhigh | 28.25 |
+| max | 63.57 |
+
+サイズが大きいほど高くなります。3840x2160 は `low` で 14.12 クレジット、`max` で 120.8 クレジットです。合計クレジット = **(出力価格 + 参照画像枚数 × 2.47〜3.71) × `n`**。参照画像 1 枚ごとに **2.47〜3.71クレジット** が加算されます。ライブバッジは単価を決める参照画像のサイズを参照できないため、この範囲を表示します。生成枚数(`n`)に応じて合計が比例します。
+
+### 画像: GPT Image 2.6 GIF
+
+**OpenAI GPT Image 2.6 GIF** はトークンではなく実行ごとに課金されます。各フレームは 1 枚の画像なので、フレームサイズ、品質、`frame_count` が価格を決めます。`frame_count` は 2〜8 フレーム、9 フレーム、10〜12 フレーム、13〜16 フレームの 4 段階で課金され、大きいサイズほど選べる段階が少なくなります。512x512・9 フレームのクレジット:
+
+| 品質 | クレジット / 実行 |
+| :----- | :---------------- |
+| low | 10.61 |
+| medium | 18.76 |
+| high | 22.72 |
+| xhigh | 40.41 |
+| max | 90.92 |
+
+384x384 で 2〜8 フレームは `low` で 3.88 クレジット、720x720 で 13〜16 フレームは `max` で 214.69 クレジットです。`frame_duration_ms` は価格に影響しません。
+
+### 画像: GPT Image 2.6 SVG
+
+**OpenAI GPT Image 2.6 SVG** は、選択したサイズと品質の静止画価格に、ベクター処理分の **0.23〜1.77クレジット** を加えた金額です。ライブバッジはベクター処理の正確な料金を決める入力を参照できないため、静止画価格にこの範囲を加えて表示します。表の範囲は静止画価格にその範囲を加えたものです。SVG のバッジに参照画像の追加料金は含まれません。サイズは 1024x1024、1024x1536、1536x1024 です。1024x1024 のクレジット:
+
+| 品質 | クレジット / 実行 |
+| :----- | :---------------- |
+| low | 7.68 – 9.22 |
+| medium | 13.36 – 14.90 |
+| high | 16.12 – 17.66 |
+| xhigh | 28.48 – 30.03 |
+| max | 63.81 – 65.35 |
+
+1024x1536 または 1536x1024 は安くなります。`low` で 6.25〜7.79 クレジット、`max` で 49.90〜51.44 クレジットです。
+
### チャット
トークンベースの課金。バッジレートは **1Kトークンあたり**。表は **100万トークンあたり**。
| モデル | 入力クレジット / 100万 | 出力クレジット / 100万 |
| :----------- | -----------------: | :------------------ |
+| gpt-6.1-sol、gpt-6-sol | 603.46 | 3017.3 |
+| gpt-6.1-terra | 301.73 | 1508.65 |
| gpt-6-astra | 3017.3 | 15086.5 |
-| gpt-6-sol | 603.46 | 3017.3 |
| gpt-6-luna | 30.17 | 150.87 |
| gpt-5.6-sol、gpt-5.6-sol-pro | 1055 | 6330 |
| gpt-5.6-terra、gpt-5.6-terra-pro | 527.5 | 3165 |
@@ -1151,10 +1299,12 @@ DALL·E と Sora は固定の画像あたりまたは秒あたりのレートを
| x-ai/grok-4.5 | 603.46 | 1810.38 |
| x-ai/grok-4.20 | 377.16 | 754.33 |
| x-ai/grok-4.3 | 377.16 | 754.33 |
+| meta/muse-wonder | 1206.92 | 6034.60 |
+| meta/muse-spark-1.3 | 377.16 | 1282.35 |
+| meta/muse-ember | 30.17 | 150.87 |
| deepseek/deepseek-v4-pro | 131.25 | 262.51 |
| deepseek/deepseek-v4-flash | 33.79 | 67.59 |
| deepseek/deepseek-v3.2 | 76.04 | 114.05 |
-| qwen/qwen3.6-max-preview | 313.80 | 1882.80 |
| qwen/qwen3.6-plus | 98.06 | 588.37 |
| qwen/qwen3.6-flash | 56.57 | 339.45 |
| mistralai/mistral-large-2512 | 150.87 | 452.60 |
@@ -1196,6 +1346,39 @@ DALL·E と Sora は固定の画像あたりまたは秒あたりのレートを
価格は概算です。1 MP のマテリアルで既定の `basecolor` と `normal` のみを有効にすると 15.09 クレジット、4 MP で 5 つのマップすべてを 4× アップスケールすると **Patina Material** が 184.06 クレジット、**Patina Material Extract** が 211.21 クレジットです。
+## Pika
+
+Pika H3 Flex は動画の秒数で課金されます。合計クレジット = **(クレジット / 秒)× 課金される秒数**。`duration` は 5〜15 秒です。`480p` と `768p` では課金される長さは `duration` と同じです。`1080p` 以上ではノードが出力をフレームグリッドに合わせるため、5 秒の生成は 5.17 秒、10 秒は 10.13 秒、15 秒は 15.08 秒として課金されます。
+
+| ノード | 解像度 | クレジット / 秒 |
+| :------------------------------ | :----- | :-------------- |
+| Pika H3 Flex Video | 480p | 3.02 |
+| Pika H3 Flex Video | 768p | 6.03 |
+| Pika H3 Flex Video | 1080p | 10.56 |
+| Pika H3 Flex Video | 2k | 12.07 |
+| Pika H3 Flex Video | 4k | 24.14 |
+| Pika H3 Flex Reference to Video | 480p | 9.05 |
+| Pika H3 Flex Reference to Video | 768p | 18.10 |
+| Pika H3 Flex Reference to Video | 1080p | 22.63 |
+| Pika H3 Flex Reference to Video | 2k | 24.14 |
+| Pika H3 Flex Reference to Video | 4k | 30.17 |
+
+**Pika H3 Flex Video** を 480p で 5 秒生成すると 15.09 クレジット、1080p で 5 秒生成すると 54.56 クレジットです。
+
+**Pika H3 Flex Reference to Video** は接続した参照動画も同じレートで課金します。各クリップは 2〜15 秒で、参照動画の合計は 15 秒までです。そのため料金は接続する参照動画の長さで変わります。480p で 5 秒生成し 2 秒の参照動画を 1 本接続すると 63.36 クレジットです。4k で 5 秒生成し参照動画を合計 15 秒接続すると 608.49 クレジットです。Pika の算例はすべて未丸めの毎秒レートで計算しています。
+
+### 音声: Pika
+
+Pika の音声ノードも秒単位で課金されます。**Pika Soundtrack** は入力動画の長さ(最大 10 分)、**Pika SFX** は `duration`(1〜20 秒)、**Pika Music** は曲の長さで課金されます。`text to music` では `duration`(10〜360 秒)、`cover` では元音声の長さ(5〜360 秒に制限)を使います。
+
+| ノード | クレジット / 秒 |
+| :-------------- | :-------------- |
+| Pika Soundtrack | 1.509 |
+| Pika SFX | 0.060 |
+| Pika Music | 0.075 |
+
+10 分の Soundtrack は 905.2 クレジット、20 秒の SFX は 1.21 クレジット、3 分の Music は 13.58 クレジットです。
+
## Pixverse
モデル **v3.5**(APIで固定。選択不可)。
@@ -1222,19 +1405,12 @@ PixVerse V6 ノードは、1〜15 秒の動画をオプションのネイティ
| PixVerse V6 | 540p | 7.18 | 9.23 |
| PixVerse V6 | 720p | 9.23 | 12.31 |
| PixVerse V6 | 1080p | 18.47 | 23.60 |
-| PixVerse V6 Max | 360p | 6.71 | 9.39 |
-| PixVerse V6 Max | 540p | 9.39 | 12.07 |
-| PixVerse V6 Max | 720p | 12.07 | 16.09 |
-| PixVerse V6 Max | 1080p | 24.14 | 30.84 |
-
-**PixVerse V6 Max** は **2026 年 11 月 3 日**より前は上記の料金が導入価格として適用されます。同日以降の秒間レートは次のとおりです:
+| PixVerse V6 Max | 360p | 8.05 | 12.07 |
+| PixVerse V6 Max | 540p | 12.07 | 14.75 |
+| PixVerse V6 Max | 720p | 14.75 | 20.12 |
+| PixVerse V6 Max | 1080p | 29.50 | 37.55 |
-| 品質 | オーディオなし | オーディオあり |
-| :---- | :------------- | :------------- |
-| 360p | 8.05 | 12.07 |
-| 540p | 12.07 | 14.75 |
-| 720p | 14.75 | 20.12 |
-| 1080p | 29.50 | 37.55 |
+**PixVerse V6 Max** の導入価格は **2026 年 11 月 2 日**まででした。上表は **2026 年 11 月 3 日**以降に適用される通常料金です。
**PixVerse V6 Fusion** は、出力時間ではなく、参照動画の 1 秒ごとに秒間レートの 2 倍が請求されます。Fusion は `PixVerse V6` のレートのみを使用します。
@@ -1242,21 +1418,10 @@ PixVerse V6 ノードは、1〜15 秒の動画をオプションのネイティ
### 画像
-終了した **arrow-1.1** モデルは実行ごとに固定料金がかかっていました:
-
- **非推奨**: Quiver は **2026年10月16日** に Arrow 1 シリーズモデルを終了し、Text to SVG と Image to SVG ノードでは `arrow-1.1`、`arrow-1.1-max`、`arrow-preview` を選択できなくなりました。代わりに Arrow 2 モデルを使用してください。以下の料金は参考として残しています。
+ **終了**: Quiver は **2026年10月16日** に Arrow 1 シリーズモデル(`arrow-1.1`、`arrow-1.1-max`、`arrow-preview`)を終了し、Text to SVG と Image to SVG ノードでは選択できなくなりました。以下に示す Arrow 2 モデルを使用してください。
-| モデル | ノード | クレジット / 実行 |
-| :------------ | :------------------ | :---------------- |
-| arrow-1.1 | Quiver Text to SVG | 60.35 |
-| arrow-1.1 | Quiver Image to SVG | 45.26 |
-| arrow-1.1-max | Quiver Text to SVG | 75.43 |
-| arrow-1.1-max | Quiver Image to SVG | 60.35 |
-| arrow-preview | Quiver Text to SVG | 90.52 |
-| arrow-preview | Quiver Image to SVG | 90.52 |
-
**arrow-2** と **arrow-2-telos** は、リクエストが消費したトークンに応じて課金されます:
| モデル | 入力クレジット / 100万トークン | 出力クレジット / 100万トークン |
@@ -1377,12 +1542,15 @@ Gen4 Turbo:**5秒** または **10秒**。合計クレジット = **(クレ
## Sonilo
-Text to Music: 合計クレジット = **(クレジット / 秒)× `duration`**(1~360秒)。
+Video to Music は入力動画の秒数で課金されます(最大 6 分)。Text to Music と Music 1.5 は音声の秒数で課金されます:合計クレジット = **(クレジット / 秒)× `duration`**。
| ノード | クレジット |
| :-------------------- | :----------- |
-| Sonilo Video to Music | 1.9 / 秒 |
-| Sonilo Text to Music | 0.5275 / 秒 |
+| Sonilo Video to Music | 2.72 / 秒 |
+| Sonilo Text to Music | 0.6789 / 秒 |
+| Sonilo Music 1.5 | 0.6789 / 秒 |
+
+Text to Music の長さは 1~360 秒です。Music 1.5 は `auto` で歌詞に合わせた長さにするか、`custom` で 10~360 秒を指定できます。
## sync.so
@@ -1453,7 +1621,7 @@ Text to Music: 合計クレジット = **(クレジット / 秒)× `duration
| :-------- | :------------- |
| Reimagine | 67.52 / 8MP |
| Bloom 2 | 96.55 / 8MP |
-| Wonder 3.5 | 24.14 / 8MP |
+| Wonder 3.5, Wonder 4 | 24.14 / 8MP |
**Topaz Image Enhance(レガシー)** は非推奨です。実行前の固定見積もりはなく、ポーリング時にAPIから返される見積もり(APIクレジット × 0.08 USD)に基づきます。入力解像度による一般的な料金:
@@ -1614,6 +1782,19 @@ Pruna P-Video-2 は生成動画の秒あたりで課金されます。合計ク
| | 1080p | オフ | 15.09 |
| | 1080p | オン | 9.05 |
+### ビデオ:P-Video Upscale
+
+Pruna P-Video Upscale は出力動画の秒あたりで課金されます。合計クレジット = **(クレジット / 秒) × 出力時間**。レートは `target` によって変わり、30 fps を超える出力は 2 倍のレートになります。ソースがターゲットより大きい場合は元のサイズのまま出力されます。
+
+`720p`、`1080p`、`2K`、`4K` には 2026 年 11 月 12 日(17:00 UTC)まで、1 秒あたり 1.51、3.02、6.03、12.07 クレジットの導入レートが適用されました。以下の表は標準レートで、同日から適用されます。
+
+| target | ターゲットサイズ | クレジット / 秒 |
+| :----- | :--------------- | :-------------- |
+| 720p | 約 0.9 MP | 3.02 |
+| 1080p | 約 2.1 MP | 6.03 |
+| 2K | 約 3.7 MP | 12.07 |
+| 4K | 約 8.3 MP | 24.14 |
+
## VEED
| ノード | オプション | クレジット |
@@ -1703,23 +1884,25 @@ Pruna P-Video-2 は生成動画の秒あたりで課金されます。合計ク
| wan2.5-t2i-preview, wan2.5-i2i-preview | — | 6.33 / 実行 |
| wan2.7-t2v, wan2.7-i2v | 720P | 21.1 / 秒 |
| | 1080P | 31.65 / 秒 |
-| wan2.7-i2v(continuation) | 720P | 21.1 / 秒(範囲) |
-| | 1080P | 31.65 / 秒(範囲) |
+| wan2.7-i2v(continuation) | 720P | 21.1 / 秒(出力 + 入力2〜5秒) |
+| | 1080P | 31.65 / 秒(出力 + 入力2〜5秒) |
| wan2.7-videoedit | 720P | 21.1 / 秒(入出力) |
| | 1080P | 31.65 / 秒(入出力) |
-| wan2.7-r2v | 720P | 21.1 / 秒(範囲) |
-| | 1080P | 31.65 / 秒(範囲) |
-| wan3.0-video | 480P | 15.09 / 秒(範囲) |
-| | 720P | 30.17 / 秒(範囲) |
-| | 1080P | 60.35 / 秒(範囲) |
-| wan3.0-video-prime | 480P | 20.52 / 秒(範囲) |
-| | 720P | 42.24 / 秒(範囲) |
-| | 1080P | 84.48 / 秒(範囲) |
+| wan2.7-r2v | 720P | 21.1 / 秒(出力のみ〜出力 + 入力5秒) |
+| | 1080P | 31.65 / 秒(出力のみ〜出力 + 入力5秒) |
+| wan3.0-video | 480P | 15.09 / 秒 |
+| | 720P | 30.17 / 秒 |
+| | 1080P | 60.35 / 秒 |
+| wan3.0-video-prime | 480P | 20.52 / 秒 |
+| | 720P | 42.24 / 秒 |
+| | 1080P | 84.48 / 秒 |
| wan2.5-t2v-preview, wan2.5-i2v-preview | 480p | 10.55 / 秒 |
| wan2.5-t2v-preview, wan2.5-i2v-preview, wan2.6-t2v, wan2.6-i2v | 720p | 21.1 / 秒 |
| wan2.5-t2v-preview, wan2.5-i2v-preview, wan2.6-t2v, wan2.6-i2v | 1080p | 31.65 / 秒 |
-| wan2.6-r2v | 720p | 21.1 / 秒(範囲) |
-| | 1080p | 31.65 / 秒(範囲) |
+| wan2.6-r2v | 720p | 21.1 / 秒(出力 + 入力2〜5秒) |
+| | 1080p | 31.65 / 秒(出力 + 入力2〜5秒) |
+
+**Wan 2.6 Reference to Video** と **Wan 2.7 continuation** は、出力時間に入力の **2〜5秒** を加えた秒数を、表と同じ秒単価で課金します。**Wan 2.7 Reference to Video** は、出力時間のみから、出力時間に入力 **5秒** を加えた範囲です。**Wan 2.7 Edit** は入力と出力を表の単価で課金します。**Wan 3.0** は表の秒単価で課金され、`duration` が `auto` のときはその秒単価が表示され、それ以外では選択した時間から `min(duration + 15, 30)` 秒までの範囲になります。
## WaveSpeed
@@ -1744,17 +1927,23 @@ Pruna P-Video-2 は生成動画の秒あたりで課金されます。合計ク
| :------------------------- | :--------- | :------- | :------------------------------------------- |
| grok-imagine-image-2.0 | 1K | generate | 12.66(中)/ 8.44(低)出力画像あたり |
| | 2K | generate | 16.88(中)/ 12.66(低)出力画像あたり |
-| grok-imagine-image-quality | 1K | generate | 10.55 / 出力画像 |
-| | 2K | generate | 14.77 / 出力画像 |
-| grok-imagine-image-pro | — | generate | 14.77 / 出力画像 |
+| grok-imagine-image-quality | 1K | generate | 8.44 / 出力画像 |
+| | 2K | generate | 12.66 / 出力画像 |
+| grok-imagine-image-pro | 1K | generate | 8.44 / 出力画像 |
+| | 2K | generate | 12.66 / 出力画像 |
| grok-imagine-image | — | generate | 4.22 / 出力画像 |
-| grok-imagine-image-2.0 | 1K | edit | 2.11 入力 + 12.66(中)/ 8.44(低)出力画像あたり |
-| | 2K | edit | 2.11 入力 + 16.88(中)/ 12.66(低)出力画像あたり |
-| grok-imagine-image-quality | 1K | edit | 2.11 入力 + 10.55 / 出力画像 |
-| | 2K | edit | 2.11 入力 + 14.77 / 出力画像 |
-| grok-imagine-image-pro | — | edit | 2.11 入力 + 14.77 / 出力画像 |
+| grok-imagine-image-2.0 | 1K | edit | 2.11–6.33 入力 + 12.66(中)/ 8.44(低)出力画像あたり |
+| | 2K | edit | 2.11–6.33 入力 + 16.88(中)/ 12.66(低)出力画像あたり |
+| grok-imagine-image-quality | 1K | edit | 2.11–6.33 入力 + 8.44 / 出力画像 |
+| | 2K | edit | 2.11–6.33 入力 + 12.66 / 出力画像 |
+| grok-imagine-image-pro | 1K | edit | 2.11 入力 + 8.44 / 出力画像 |
+| | 2K | edit | 2.11 入力 + 12.66 / 出力画像 |
| grok-imagine-image | — | edit | 0.422–1.266 入力 + 4.22 / 出力画像 |
+非 pro の edit モデルでは、入力画像 1 枚あたり **2.11クレジット**、入力は **1〜3枚**(**2.11–6.33**)です。`grok-imagine-image` は入力画像 1 枚あたり **0.422クレジット**(**0.422–1.266**)です。pro の edit 行は入力画像 1 枚分の **2.11クレジット** だけを課金します。
+
+`grok-imagine-image-quality` と `grok-imagine-image-pro` は xAI により廃止されました。画像ノードはこの 2 つのモデルを引き続き表示し、`grok-imagine-image-2.0` の低品質レートで課金します。
+
### 動画
| モデル | 解像度 | クレジット |
@@ -1772,6 +1961,8 @@ Pruna P-Video-2 は生成動画の秒あたりで課金されます。合計ク
| | 720p | 21.12 / 秒 出力 + 0.603 / 参照画像 |
| grok-imagine-video(extend)| — | 6.03–45.26 入力 + 15.09 × 延長時間 秒 出力 / 実行 |
+**Extend** は延長 1 秒あたり **15.09クレジット** に加え、ソース動画の入力分として 2 秒なら **6.03クレジット**、15 秒なら **45.26クレジット** を課金します(ソース動画は 2〜15 秒である必要があります)。ライブバッジはソースの長さを参照できないため、この **6.03–45.26** の入力範囲に、延長秒数分の料金を加えて表示します。
+
## クラウド GPU
| gpu_type | クレジット |
diff --git a/ko/tutorials/partner-nodes/pricing.mdx b/ko/tutorials/partner-nodes/pricing.mdx
index ed648b0231..2f2a1fbe47 100644
--- a/ko/tutorials/partner-nodes/pricing.mdx
+++ b/ko/tutorials/partner-nodes/pricing.mdx
@@ -3,54 +3,56 @@ title: "Partner Node 가격"
description: "Anthropic, OpenAI, Kling, Luma 등 모든 제공업체의 Partner Node 크레딧 가격을 모델, 토큰 유형, 생성 유형별로 구분합니다."
sidebarTitle: "가격 정책"
mode: "wide"
-translationSourceHash: 2ef8ce31
+translationSourceHash: f68d51fd
translationFrom: tutorials/partner-nodes/pricing.mdx
translationBlockHashes:
"_intro": 1ec20604
- "Anthropic": 61f55fec
+ "Anthropic": 1757fac0
"Beeble": d1b19f61
- "BFL": 7ae11dc6
+ "BFL": 0a63f1b4
"Bria": 94080005
- "ByteDance": d2af4769
+ "Breeze": f38692c6
+ "ByteDance": fec72d04
"Clarity AI": 3d568b52
"ElevenLabs": cadb48cc
- "FishAudio": c957b7da
+ "FishAudio": 24ec3ea0
"Magnific": 4000e7e1
- "Google": b8be5b63
+ "Google": 958913c2
"HappyHorse": bceb584a
- "HeyGen": 7f08ab4e
+ "HeyGen": 248354a7
"Higgsfield": eed76dcb
"Hitpaw": 320c0537
- "Ideogram": da793624
+ "Ideogram": de0981df
"Krea": bcad7ba2
- "Kling": 32a38b1d
- "Lightricks": 3ba8d019
- "Luma": 9f809a47
+ "Kling": db91f830
+ "Lightricks": bbcb475e
+ "Luma": a2f2a8de
"Meshy": c057a289
"Meta": a253fe9a
"Mirelo": 2055c18a
- "Minimax": 07381eba
- "OpenAI": 17e24f34
- "OpenRouter": ef37d425
+ "Minimax": 3f9b84c5
+ "OpenAI": e51a94f8
+ "OpenRouter": 3de75ece
"Patina": d489c37e
- "Pixverse": bdb9b1c9
- "Quiver": 5538ce9c
+ "Pika": 66ba8620
+ "Pixverse": 86f21b3f
+ "Quiver": f84ffa99
"Qwen": 22dbe478
"Recraft": fc591c24
"Reve": 02eb39a9
"Rodin 3D": 2eb41834
"Runway": bd14d7ff
- "Sonilo": 48b9c581
+ "Sonilo": 66cc2806
"sync.so": 590402a9
"Tencent": cbe22aa6
- "Topaz": 5d08cf76
+ "Topaz": f557fb17
"Tripo": 842e861e
- "Pruna": f1b6a267
+ "Pruna": 08c63f88
"VEED": 1818bad3
"Vidu": be19d9ca
"Wan": 99df0ce3
"WaveSpeed": 349a7c9d
- "xAI": 28ff86de
+ "xAI": 6efd8650
"Cloud GPU": e228bb68
---
@@ -79,6 +81,7 @@ translationBlockHashes:
| :--------- | -----------------: | --------------------------: | --------------------------: | ---------------------: | :------------------ |
| Opus 5.5 | 1.207 | 1.509 | 2.414 | 0.121 | 6.035 |
| Opus 5, Opus 4.8 | 1.509 | 1.886 | 3.017 | 0.151 | 7.543 |
+| Fable 5.2 | 2.414 | 3.017 | 4.828 | 0.241 | 12.069 |
| Fable 5.1, Fable 5 | 3.017 | 3.772 | 6.035 | 0.302 | 15.087 |
| Sonnet 5.5, Sonnet 5 | 0.603 | 0.754 | 1.207 | 0.060 | 3.017 |
@@ -104,17 +107,19 @@ translationBlockHashes:
### 이미지
-| 노드 | 크레딧 |
-| :------------------------- | :---------------------------: |
-| FLUX 1.1 [pro] Ultra 이미지 | 12.66 / 실행 |
-| FLUX.1 Kontext [pro] 이미지 | 8.44 / 실행 |
-| FLUX.1 Kontext [max] 이미지 | 16.88 / 실행 |
-| FLUX.1 이미지 확장 | 10.55 / 실행 |
-| FLUX.1 이미지 채우기 | 10.55 / 실행 |
-| FLUX 이미지 지우기 | 6.33 / 실행 + 0.84 / 추가 MP |
-| FLUX 가상 피팅-On | 7.91 / 실행 + 1.06 / 추가 MP |
-| FLUX.2 [pro] | 6.33 / 실행 + 3.17 / 추가 MP |
-| FLUX.2 [max] | 14.77 / 실행 + 6.33 / 추가 MP |
+| 노드 | 크레딧 | 크레딧(참조 이미지 포함) |
+| :-------------------------- | :---------------------------: | :-----------------------------------: |
+| FLUX 1.1 [pro] Ultra 이미지 | 12.66 / 실행 | — |
+| FLUX.1 Kontext [pro] 이미지 | 8.44 / 실행 | — |
+| FLUX.1 Kontext [max] 이미지 | 16.88 / 실행 | — |
+| FLUX.1 이미지 확장 | 10.55 / 실행 | — |
+| FLUX.1 이미지 채우기 | 10.55 / 실행 | — |
+| FLUX 이미지 지우기 | 6.33 / 실행 + 0.84 / 추가 MP | — |
+| FLUX 가상 피팅-On | 7.91 / 실행 + 1.06 / 추가 MP | — |
+| FLUX.2 [pro] | 6.33 / 실행 + 3.17 / 추가 MP | 9.50 – 31.65 / 실행 + 3.17 / 추가 MP |
+| FLUX.2 [max] | 14.77 / 실행 + 6.33 / 추가 MP | 21.10 – 65.41 / 실행 + 6.33 / 추가 MP |
+
+참조 이미지를 연결하면 세 번째 열의 요금이 적용됩니다. 텍스트만 사용할 때 **FLUX.2 [pro]**는 **6.33 크레딧**에 추가 출력 메가픽셀(`ceil(W × H / 1 MP) − 1`)당 **3.17 크레딧**을 더한 금액입니다. **FLUX.2 [max]**는 **14.77 크레딧**에 추가 출력 메가픽셀당 **6.33 크레딧**을 더한 금액입니다. 참조 이미지가 있으면 입력 메가픽셀당 [pro]는 **3.17 크레딧**, [max]는 **6.33 크레딧**이 추가됩니다(각 참조 이미지의 `ceil(W × H / 1 MP)`를 합산, 최대 8장). 세 번째 열은 출력이 1 MP이고 참조 입력이 1–8 MP일 때의 범위(**9.50–31.65** / **21.10–65.41**)이며, 추가 출력 메가픽셀은 텍스트 전용 열과 같은 단가로 더해집니다. [pro]의 **3.17**은 반올림한 단가이며, 공개된 **9.50–31.65** 구간은 참조 입력 1–8 MP일 때의 라이브 배지 끝값입니다.
### 이미지: FLUX 3 Image
@@ -149,14 +154,14 @@ translationBlockHashes:
### 비디오: FLUX 3 Max
-총 크레딧 = **(크레딧 / 초) × `duration`** (`duration`은 5–20초). `reasoning_level`과 `generate_audio`는 가격에 영향을 주지 않습니다. 출시 프로모션 요금은 **2026년 11월 7일(00:00 UTC)**까지 적용되며, 이후에는 표준 요금이 적용됩니다.
+총 크레딧 = **(크레딧 / 초) × `duration`** (`duration`은 5–20초). `reasoning_level`과 `generate_audio`는 가격에 영향을 주지 않습니다. 720p 12.82, 1080p 21.88 크레딧 / 초의 출시 요금은 2026년 11월 7일(00:00 UTC)까지 적용되었습니다. 아래 표는 표준 요금이며 같은 시점부터 적용됩니다.
-| 노드 | 해상도 | 프로모션 크레딧 / 초 | 표준 크레딧 / 초 |
-| :------------------------ | :---- | :----------- | :--------- |
-| FLUX 3 Max Text to Video | 720p | 12.82 | 51.29 |
-| FLUX 3 Max Text to Video | 1080p | 21.88 | 87.50 |
-| FLUX 3 Max Image to Video | 720p | 12.82 | 51.29 |
-| FLUX 3 Max Image to Video | 1080p | 21.88 | 87.50 |
+| 노드 | 해상도 | 크레딧 / 초 |
+| :------------------------ | :---- | :------- |
+| FLUX 3 Max Text to Video | 720p | 51.29 |
+| FLUX 3 Max Text to Video | 1080p | 87.50 |
+| FLUX 3 Max Image to Video | 720p | 51.29 |
+| FLUX 3 Max Image to Video | 1080p | 87.50 |
### 비디오: Flux Video Edit
@@ -172,10 +177,10 @@ translationBlockHashes:
| 노드 | 모드 | 크레딧 / MP / 초 |
| :--------------------- | :------- | :-------------- |
-| FLUX 비디오 업스케일 | creative | 15.09 |
-| FLUX 비디오 업스케일 | precise | 10.56 |
+| FLUX 비디오 업스케일 | creative | 14.39 |
+| FLUX 비디오 업스케일 | precise | 10.07 |
-1080p 출력(2.07MP)은 creative 모드에서 초당 31.24 크레딧, precise 모드에서 21.86 크레딧입니다. 4K 출력(8.29MP)은 각각 125.10와 87.54입니다.
+1080p 출력(2.07MP)은 creative 모드에서 초당 29.83 크레딧, precise 모드에서 20.88 크레딧입니다. 4K 출력(8.29MP)은 각각 119.34와 83.54입니다.
## Bria
@@ -212,6 +217,18 @@ translationBlockHashes:
| Bria 비디오 배경 제거 (투명) | 10.55 / 초 |
| Bria 비디오 지우개 | 13.58 / 초 |
+## Breeze
+
+Breeze 음성 노드는 모두 **9.05 크레딧 / 1K 과금 문자**로 청구되며, CJK 문자는 두 글자로 계산합니다. 텍스트 음성 변환과 보이스 디자인은 입력한 텍스트로 청구됩니다. 보이스 복제, 리믹스, 현지화도 같은 단가이지만, 문자 수는 실행이 끝난 뒤에야 알 수 있습니다. 리믹스와 현지화는 생성된 미리보기 텍스트, 복제는 업체가 보고한 과금 단위입니다. 그래서 이 세 노드의 배지는 대략 **0.66–2.66**, **0.72–1.52**, **0.76–2.29** 크레딧 범위를 보여줍니다.
+
+| 노드 | 크레딧 |
+| :---------------------- | :-------------- |
+| Breeze 텍스트 음성 변환 | 9.05 / 1K chars |
+| Breeze 보이스 디자인 | 9.05 / 1K chars |
+| Breeze 보이스 복제 | 9.05 / 1K chars |
+| Breeze 보이스 리믹스 | 9.05 / 1K chars |
+| Breeze 보이스 현지화 | 9.05 / 1K chars |
+
## ByteDance
### LLM: ByteDance Seed
@@ -232,15 +249,17 @@ translationBlockHashes:
| 모델 | 크레딧 / 이미지 |
| :-------------------------------------------------- | :------------------- |
+| seedream 5.5 (`seedream-5-5-261028`) | 6.03 (1K) / 10.56 (2K) |
| seedream 5.0 pro (`seedream-5-0-pro-260628`) | 9.50 (1K) / 18.99 (2K) |
| seedream 5.0 flash (`seedream-5-0-flash-260915`) | 5.43 |
| seedream 5.0 lite (`seedream-5-0-260128`) | 7.39 |
| seedream-4-5-251128 | 8.44 |
| seedream-4-0-250828 | 6.33 |
+| Seedream 5.5 레이어 분리 | 6.03 (1K, 1.5K) / 10.56 (2K) |
| Seedream 5.0 Pro 레이어 분리 | 6.79 (1K, 1.5K) / 13.58 (2K) |
| Seedream 5.0 Flash 레이어 분리 | 5.43 |
-Seedream 5.0 Pro는 해상도 기반 요금제를 사용합니다: **1K** (≤2.61 MP, **9.50 크레딧**) 및 **2K** (>2.61 MP, **18.99 크레딧**). Seedream 5.0 Flash는 해상도와 관계없이 이미지당 **5.43 크레딧**의 정액으로 청구됩니다. 각 요청의 첫 번째 이미지는 **무료**입니다. 첫 번째 이미지 이후의 추가 이미지(출력 이미지와 입력/참조 이미지 모두 포함)마다 **0.63 크레딧**이 추가됩니다.
+Seedream 5.0 Pro는 해상도 기반 요금제를 사용합니다: **1K** (≤2.61 MP, **9.50 크레딧**) 및 **2K** (>2.61 MP, **18.99 크레딧**). Seedream 5.0 Flash는 해상도와 관계없이 이미지당 **5.43 크레딧**의 정액으로 청구됩니다. 각 요청의 첫 번째 이미지는 **무료**입니다. 첫 번째 이미지 이후의 추가 이미지(출력 이미지와 입력/참조 이미지 모두 포함)마다 **0.63 크레딧**이 추가됩니다. **Seedream 5.5**는 약 2.6MP까지 이미지당 **6.03 크레딧**, 그 이상은 **10.56 크레딧**이며, 참조 이미지를 연결할 때마다 **2.41 크레딧**이 추가됩니다. **Seedream 5.5 레이어 분리**는 1K/1.5K에서 **6.03 크레딧**, 2K에서 **10.56 크레딧**이며, 해당 실행의 이미지 수에 따라 과금됩니다.
### 이미지: ByteDance 이미지 (지원 중단됨)
@@ -254,24 +273,45 @@ Seedream 5.0 Pro는 해상도 기반 요금제를 사용합니다: **1K** (≤2.
### 비디오: Seedance 2.5
-토큰 기반 과금. 예상 토큰 소비량 = **(입력 비디오 길이 + 출력 비디오 길이) × 출력 비디오 너비 × 출력 비디오 높이 × 출력 비디오 프레임레이트 ÷ 1024**. 텍스트, 이미지, 오디오만 입력한 경우 입력 비디오 길이는 0입니다. 실제 사용량은 API를 기준으로 합니다.
+토큰 기반 과금. ComfyUI 라이브 가격 배지는 토큰을 **floor((출력 너비 × 출력 높이 ÷ 1024) × (24 × 과금 길이 + 1))** 로 추정한 뒤, 아래 표의 크레딧 / 1K 토큰 요율을 곱합니다(`+ 1`은 추가 1프레임입니다). 과금 길이는 출력 길이이며, 참조 비디오가 연결되어 있으면 입력 비디오 길이를 더합니다. 텍스트, 이미지, 오디오만 입력한 경우 입력 비디오 길이는 0입니다. 실제 사용량은 API를 기준으로 합니다.
+
+| 모델 | 해상도 | 비디오 입력 | 크레딧 / 1K 토큰 |
+| :----------------- | :--------- | :---------- | ---------------: |
+| Seedance 2.5 | 480p, 720p | 아니요 | 3.229 |
+| Seedance 2.5 | 480p, 720p | 예 | 1.931 |
+| Seedance 2.5 | 1080p | 아니요 | 3.530 |
+| Seedance 2.5 | 1080p | 예 | 2.112 |
+| Seedance 2.5 | 4k | 아니요 | 1.418 |
+| Seedance 2.5 | 4k | 예 | 0.875 |
+| Seedance 2.5 Flash | 480p, 720p | 아니요 | 2.112 |
+| Seedance 2.5 Flash | 480p, 720p | 예 | 1.297 |
+
+**비디오 입력**은 참조 비디오가 연결됨을 의미합니다. 텍스트, 이미지 및 오디오 참조는 비디오 입력 없음으로 간주됩니다. 라이브 배지는 위의 계산식 외에 별도의 최소 토큰 과금 하한을 적용하지 않습니다.
+
+**Seedance 2.5 Flash**는 480p와 720p만, 최대 30초까지 생성하며 두 해상도의 요율이 같습니다. 비디오 입력이 없는 5초 720p 클립은 약 230 크레딧입니다.
+
+같은 길이에서 4k 프레임은 1080p 프레임보다 토큰이 약 4배 많으므로 1K 토큰당 단가는 낮지만 총 비용은 더 높습니다.
+
+참조 비디오는 최소 2초 이상이어야 하며, 입력 합계는 최대 30초입니다. 위의 토큰 계산식으로 구한 16:9 출력(24fps)의 예상 요금:
-| 모델 | 해상도 | 비디오 입력 | 크레딧 / 1K 토큰 |
-| :----------- | :--------- | :------- | --------------: |
-| Seedance 2.5 | 480p, 720p | 아니요 | 3.229 |
-| Seedance 2.5 | 480p, 720p | 예 | 1.931 |
-| Seedance 2.5 | 1080p | 아니요 | 3.530 |
-| Seedance 2.5 | 1080p | 예 | 2.112 |
+| 해상도 | 5초 출력, 비디오 입력 없음 | 5초 출력 + 2초 비디오 입력 | 5초 출력 + 30초 비디오 입력 |
+| :--------- | -----------------------: | -----------------------: | ------------------------: |
+| 480p | 156.38 | 130.64 | 650.12 |
+| 720p | 351.58 | 293.72 | 1461.63 |
+| 1080p | 865.00 | 722.82 | 3596.98 |
+| 4k | 1389.91 | 1197.81 | 5960.70 |
-**비디오 입력**은 참조 비디오가 연결됨을 의미합니다. 텍스트, 이미지 및 오디오 참조는 비디오 입력 없음으로 간주됩니다. 입력에 비디오가 포함된 경우 최소 토큰 소비량 제한이 적용됩니다: 예상 소비량이 최소값보다 낮으면 최소 소비량으로 청구됩니다.
+참조 비디오를 연결하면 1K 토큰당 요금이 낮아지므로, 5초 출력에 2초 참조를 더한 요금이 비디오 입력 없는 5초보다 저렴할 수 있습니다. 라이브 배지는 입력 비디오의 길이를 알 수 없으므로, 참조 비디오가 연결된 5초 실행에서는 ‘출력 5초만(비디오 입력 요금)’부터 ‘출력 5초 + 입력 30초’까지의 범위를 표시합니다.
-참조 비디오는 최소 2초 이상이어야 하며, 입력 합계는 최대 30초입니다. 16:9 출력(24fps)의 예상 요금:
+**Seedance 2.5 Draft** 모델로 실행하면 위 첫 번째 표의 480p 토큰 요금으로 480p 미리보기를 생성하고 `draft_task_id`를 반환합니다. **ByteDance Seedance 2.5 Draft to Final Video**는 그 드래프트를 1080p 또는 4k 최종 영상으로 렌더링합니다. 최종 영상은 드래프트의 프롬프트, 참조, 길이, 오디오 설정을 그대로 사용합니다. 이 노드의 라이브 배지는 같은 계산식으로 길이에 의존하지 않는 크레딧 범위를 표시합니다(하한은 참조 영상 없는 4초 16:9 최종 영상, 상한은 30초 최종 영상에 30초 참조 영상을 더한 경우). 상한은 1080p에서 21:9, 4k에서 4:3 또는 3:4입니다:
+
+| 노드 | 해상도 | 크레딧 |
+| :------------------------------------------ | :----- | ----------------: |
+| ByteDance Seedance 2.5 Draft to Final Video | 1080p | 693.43 – 6202.64 |
+| ByteDance Seedance 2.5 Draft to Final Video | 4k | 1114.23 – 10214.09 |
+
+참조 영상이 없는 4초 1080p·16:9 최종 영상은 693.43 크레딧, 30초 참조 영상이 있는 30초 1080p·16:9 최종 영상은 6163.19 크레딧입니다. 두 값 모두 배지의 추가 1프레임을 포함합니다. 화면 비율이 달라지면 요금도 조금 달라집니다. 1080p 상한 6202.64는 21:9, 4k 상한 10214.09는 4:3 또는 3:4인 경우입니다. 최종 청구는 실제 API 사용량을 따릅니다.
-| 해상도 | 5초 출력, 비디오 입력 없음 | 5초 출력 + 2–30초 비디오 입력 |
-| :--------- | -----------------------: | ---------------------------: |
-| 480p | 155.1 | 166.9 – 649.3 |
-| 720p | 348.8 | 375.4 – 1459.8 |
-| 1080p | 857.8 | 923.9 – 3592.7 |
### 비디오: Seedance 2.0
480p와 720p는 동일한 요금입니다. Fast 및 Mini는 1080p 또는 4K를 지원하지 않습니다.
@@ -311,7 +351,7 @@ Seedream 5.0 Pro는 해상도 기반 요금제를 사용합니다: **1K** (≤2.
| seedance-1-0-pro-fast-251015 | — | 211 |
- **지원 중단**: `seedance-1-5-pro-251215`는 BytePlus가 해당 모델을 종료하는 2026년 11월 11일에 서비스가 종료되며, ComfyUI도 같은 날 ByteDance 1.x 노드에서 이 모델을 제거합니다. 따라서 이 페이지에서도 해당 모델의 가격 행을 삭제했습니다. 대신 `seedance-1-0-pro-250528`, `seedance-1-0-pro-fast-251015` 또는 Seedance 2.x 모델을 사용하세요.
+ **지원 종료**: BytePlus가 2026년 11월 11일에 `seedance-1-5-pro-251215`를 종료하여 이 페이지에서 해당 모델의 가격 행을 삭제했습니다. 대신 `seedance-1-0-pro-250528`, `seedance-1-0-pro-fast-251015` 또는 Seedance 2.x 모델을 사용하세요.
**종료됨**: `seedance-1-0-lite-t2v-250428` 및 `seedance-1-0-lite-i2v-250428`, 그리고 **ByteDance 참조 이미지 to 비디오** 노드는 서비스가 종료되어 ComfyUI에서 제거되었습니다. 아래 가격은 참고용으로만 유지됩니다.
@@ -403,7 +443,7 @@ Seedream 5.0 Pro는 해상도 기반 요금제를 사용합니다: **1K** (≤2.
## FishAudio
-**기간 한정 혜택**: 2026년 10월 13일부터 11월 12일까지 Drama 3은 약 **3.02** 크레딧 / 1K 바이트로 청구됩니다. 표의 `drama-3` 행은 표준 가격이며 2026년 11월 13일부터 적용됩니다.
+Drama 3은 2026년 10월 13일부터 11월 12일까지 출시 요금인 약 **3.02** 크레딧 / 1K 바이트로 청구되었습니다. 표의 `drama-3` 행은 표준 가격이며 2026년 11월 13일부터 적용됩니다.
| 노드 | 모델 | 크레딧 |
@@ -446,6 +486,9 @@ Seedream 5.0 Pro는 해상도 기반 요금제를 사용합니다: **1K** (≤2.
| 모델 | 입력 크레딧 / 1K | 출력 크레딧 / 1K |
| :------------------------------------------------------ | -----------------: | :------------------ |
+| Gemini 4 Hydrogen (`gemini-4-hydrogen`) | 0.4526 | 2.263 |
+| Gemini 4 Argon (`gemini-4-argon`) | 1.20692 | 6.0346 |
+| Gemini 4 Carbon (`gemini-4-carbon`) | 1.20692 | 6.0346 |
| Gemini 3.8 Flash (`gemini-3.8-flash`) | 0.4526 | 2.263 |
| Gemini 3.7 Flash (`gemini-3.7-flash`) | 0.4526 | 2.263 |
| Gemini 3.5 Flash (`gemini-3.5-flash`) | 0.3165 | 1.899 |
@@ -477,6 +520,10 @@ Seedream 5.0 Pro는 해상도 기반 요금제를 사용합니다: **1K** (≤2.
| | 720p | 30.57 |
| | 1080p | 45.87 |
| | 4k | 91.76 |
+| Gemini Omni Flash 1.2 (`gemini-omni-1.2-flash`) | 360p | 6.98 |
+| | 720p | 20.97 |
+| | 1080p | 31.46 |
+| | 4k | 62.92 |
### 비디오: Google Veo
@@ -532,23 +579,33 @@ Seedream 5.0 Pro는 해상도 기반 요금제를 사용합니다: **1K** (≤2.
| HeyGen Video Translate | speed mode | 4.07 / 초 |
| HeyGen Text to Speech | N/A | ~0.20 / 초 |
| HeyGen Video 1.0 Reference to Video | 480p, 참조 자료 없음 | 3.02 / 초 |
-| HeyGen Video 1.0 Reference to Video | 768p, 참조 자료 없음 | 4.53 / 초 |
-| HeyGen Video 1.0 Reference to Video | 2K, 참조 자료 없음 | 13.58 / 초 |
-| HeyGen Video 1.0 Reference to Video | 480p, 참조 자료 있음 | 6.03 / 초 |
-| HeyGen Video 1.0 Reference to Video | 768p, 참조 자료 있음 | 9.05 / 초 |
-| HeyGen Video 1.0 Reference to Video | 2K, 참조 자료 있음 | 27.16 / 초 |
-| HeyGen Video 1.0 Image to Video | 480p | 3.02 / 초 |
-| HeyGen Video 1.0 Image to Video | 768p | 4.53 / 초 |
-| HeyGen Video 1.0 Image to Video | 2K | 13.58 / 초 |
+| HeyGen Video 1.0 Reference to Video | 768p, 참조 자료 없음 | 5.43 / 초 |
+| HeyGen Video 1.0 Reference to Video | 1080p, 참조 자료 없음 | 12.07 / 초 |
+| HeyGen Video 1.0 Reference to Video | 2K, 참조 자료 없음 | 15.09 / 초 |
+| HeyGen Video 1.0 Reference to Video | 480p, 참조 자료 있음 | 10.56 / 초 |
+| HeyGen Video 1.0 Reference to Video | 768p, 참조 자료 있음 | 15.09 / 초 |
+| HeyGen Video 1.0 Reference to Video | 1080p 또는 2K, 참조 자료 있음 | 27.16 / 초 |
+| HeyGen Video 1.0 Reference to Video | 세 번째부터의 추가 참조 이미지 | 6.03 / 장 |
+| HeyGen Video 1.0 Image to Video | 480p | 3.02 / 초 |
+| HeyGen Video 1.0 Image to Video | 768p | 5.43 / 초 |
+| HeyGen Video 1.0 Image to Video | 1080p | 12.07 / 초 |
+| HeyGen Video 1.0 Image to Video | 2K | 15.09 / 초 |
+| HeyGen Video 1.0 Image to Video | 15초 초과, 480p | 13.58 / 초 |
+| HeyGen Video 1.0 Image to Video | 15초 초과, 768p | 22.63 / 초 |
+| HeyGen Video 1.0 Image to Video | 15초 초과, 1080p 또는 2K | 31.68 / 초 |
| HeyGen Voice 1.0 Instant Clone | N/A | 무료 |
| HeyGen Voice 1.0 Text to Speech | `heygen-voice-1` | 9.05 / 1K자 |
| HeyGen Voice 1.0 Text to Speech | `heygen-voice-1-turbo` | 2.26 / 1K자 |
-**텍스트 음성 변환 가격**: 2026년 10월 31일까지 1K자당 약 **4.53** 크레딧(`heygen-voice-1`), **1.21** 크레딧(`heygen-voice-1-turbo`)의 출시 요금이 적용되었습니다. 위 표는 표준 가격이며 2026년 11월 1일부터 적용됩니다.
+**텍스트 음성 변환 가격**: 2026년 11월 1일(00:00 UTC)까지 1K자당 약 **4.53** 크레딧(`heygen-voice-1`), **1.21** 크레딧(`heygen-voice-1-turbo`)의 출시 요금이 적용되었습니다. 위 표는 표준 가격이며 그 시점부터 적용됩니다.
아바타 비디오는 선택한 아바타 룩의 유형(Digital Twin / Photo Avatar / Studio Avatar) 요율로 초 단위로 청구됩니다. **자동**은 룩에 따라 엔진을 자동 선택하므로 요율이 위 표의 전체 범위에 걸칩니다. 사용자 지정 아바타 ID는 해당 유형에 따라 청구됩니다.
-HeyGen Video 1.0 클립은 5~15초이며, 선택한 해상도의 요율로 청구됩니다. Reference to Video는 참조 이미지나 참조 동영상을 연결하기 전까지 참조 자료 없는 요율로 청구되며, 참조 동영상마다 청구 시간이 5초 늘어납니다. **2K** 출력에는 `16:9` 또는 `9:16` 종횡비가 필요하며, `auto`는 참조 미디어가 없을 때만 `2K`에서 작동합니다.
+**기간 한정**: 50% 할인 프로모션 요율은 2026년 12월 1일(00:00 UTC)까지 적용됩니다. 위 표의 Video 1.0 행은 현재 프로모션 요율입니다. 장편 클립 요율은 이미 표준 가격입니다. 2026년 12월 1일(00:00 UTC)부터는 표준 요율이 적용됩니다. 참조 자료 없음 및 Image to Video(15초 이하)는 초당 **6.03 / 10.86 / 24.14 / 30.17** 크레딧, 참조 자료 있음은 초당 **21.12 / 30.17 / 54.31** 크레딧, 세 번째부터의 추가 참조 이미지는 장당 **12.07 크레딧**입니다. 장편 클립 요율은 그대로입니다.
+
+HeyGen Video 1.0 클립은 5~30초이며 초 단위로 청구됩니다. 15초를 넘는 클립은 두 노드 모두 참조 자료가 없을 때 장편 행의 요율로 청구됩니다. 참조 자료를 연결한 경우에는 15초 초과를 사용할 수 없습니다. 이미지든 비디오든 하나라도 연결하면, 처음 두 장에 장당 추가 요금이 없더라도 **Reference to Video**는 참조 자료 있음 요율로 전환됩니다. 처음 두 장의 참조 이미지는 무료이고, 세 번째부터는 장당 **6.03 크레딧**이 추가됩니다(이 추가는 2026년 12월 1일(00:00 UTC)부터 **12.07 크레딧**이 됩니다). 참조 영상은 총 15초까지 참조 자료 있음 초당 요율로 청구됩니다.
+
+`auto_duration`은 `duration` 대신 프롬프트의 대사 길이를 사용합니다(8~30초, 참조 자료를 연결한 경우 8~15초에 참조 영상 길이가 더해집니다). **1080p**와 **2K** 출력에는 `16:9` 또는 `9:16` 종횡비가 필요하고, `auto`는 참조 미디어가 없을 때만 해당 해상도에서 작동합니다.
## Higgsfield
@@ -659,15 +716,15 @@ HeyGen Video 1.0 클립은 5~15초이며, 선택한 해상도의 요율로 청
## Ideogram
-| 노드 | 터보 | 캐릭터 참조 | 렌더링 속도 | 크레딧 / 이미지 |
-| :---------- | :---- | :------------------ | :-------------- | :-------------- |
-| Ideogram V1 | 거짓 | — | — | 18.1 |
-| Ideogram V1 | 참 | — | — | 6.03 |
-| Ideogram V2 | 거짓 | — | — | 24.14 |
-| Ideogram V2 | 참 | — | — | 15.09 |
-| Ideogram V4 | — | — | TURBO | 9.05 |
-| Ideogram V4 | — | — | DEFAULT | 18.1 |
-| Ideogram V4 | — | — | QUALITY | 30.17 |
+
+ **지원 중단**: ComfyUI에서 **Ideogram V1** 및 **Ideogram V2** 노드가 제거되어 이 페이지에는 해당 가격이 더 이상 표시되지 않습니다. 대신 **Ideogram V4**를 사용하세요.
+
+
+| 노드 | 렌더링 속도 | 크레딧 / 이미지 |
+| :---------- | :---------- | :-------------: |
+| Ideogram V4 | TURBO | 9.05 |
+| Ideogram V4 | DEFAULT | 18.1 |
+| Ideogram V4 | QUALITY | 30.17 |
| 노드 | 품질 | 해상도 | 크레딧 / 이미지 |
| :--------------- | :------ | :----- | :-------------- |
@@ -726,13 +783,41 @@ HeyGen Video 1.0 클립은 5~15초이며, 선택한 해상도의 요율로 청
| Krea 3 | 2K | high | 24.14 | 6.03 - 21.12 |
| Krea 3 | 2K | ultra | 78.45 | 6.03 - 21.12 |
-2K 출력에서는 참조 이미지의 픽셀 크기에 따라 참조 이미지당 **6.03~21.12 크레딧**이 추가됩니다. 1K 출력에서는 참조 이미지가 약 1 MP로 다운샘플링되므로 이미지당 **6.03 크레딧**이 정액으로 추가됩니다. `ultra`는 `resolution` 값과 관계없이 항상 2K로 출력되며, 참조 이미지가 1장 이상 필요합니다.
+2K 출력에서는 참조 이미지당 **6.03–21.12 크레딧**이 추가됩니다(하한 **6.03**, 상한 **21.12**. 노드는 참조 이미지의 픽셀 크기에 따라 이 두 값 사이에서 요금을 산정합니다). 1K 출력에서는 참조 이미지가 약 1 MP로 다운샘플링되므로 이미지당 **6.03 크레딧**이 정액으로 추가됩니다. `ultra`는 `resolution` 값과 관계없이 항상 2K로 출력되며, 참조 이미지가 1장 이상 필요합니다.
## Kling
### 비디오: 크레딧 / 초
-총 크레딧 = **(크레딧 / 초) × `duration`**, 단 **Kling 3.0 Omni Edit Video** 및 **Kling 모션 컨트롤**은 예외입니다 (출력 초당 요금 부과; 편집 시 duration은 입력 비디오를 따름).
+총 크레딧 = **(크레딧 / 초) × `duration`**, 단 **Kling 3.0 Omni Edit Video**, **Kling 4.0 Edit Video** 및 **Kling 모션 컨트롤**은 예외입니다 (출력 초당 요금 부과; 편집 시 duration은 입력 비디오를 따름).
+
+#### kling-4.0
+
+`kling-4.0`은 3~30초, 720p 또는 1080p 영상을 생성합니다. `kling-4.0-flash`는 3~20초, 720p입니다.
+
+| 노드 | 모델 | 해상도 | 크레딧 / 초 | 참조 비디오 사용 시 크레딧 / 초 |
+| :--------------------------- | :-------------- | :----- | :---------- | :------------------------------ |
+| Kling 4.0 Text to Video | kling-4.0 | 720p | 47.37 | — |
+| Kling 4.0 Text to Video | kling-4.0 | 1080p | 70.91 | — |
+| Kling 4.0 Image to Video | kling-4.0 | 720p | 47.37 | — |
+| Kling 4.0 Image to Video | kling-4.0 | 1080p | 70.91 | — |
+| Kling 4.0 Reference to Video | kling-4.0 | 720p | 47.37 | 28.36 |
+| Kling 4.0 Reference to Video | kling-4.0 | 1080p | 70.91 | 42.54 |
+| Kling 4.0 Text to Video | kling-4.0-flash | 720p | 28.36 | — |
+| Kling 4.0 Image to Video | kling-4.0-flash | 720p | 28.36 | — |
+| Kling 4.0 Reference to Video | kling-4.0-flash | 720p | 28.36 | 19.01 |
+
+**Kling 4.0 Reference to Video**에서 참조 비디오를 연결하면 실행 전체가 더 낮은 요율로 바뀝니다. 출력의 매초와 참조 비디오의 매초가 모두 그 요율로 청구됩니다. 각 클립은 3~30초(`kling-4.0-flash`는 3~20초)이며, 모든 클립의 합계도 같은 상한을 넘을 수 없습니다. 5초 720p 출력에 30초 참조 비디오를 붙이면 992.7 크레딧입니다.
+
+**Kling 4.0 Edit Video**는 항상 더 낮은(참조 비디오) 요율로 청구됩니다. **크레딧 = 단가 × (입력 비디오 초 + 참조 비디오 초)**. 출력은 입력과 같은 길이이므로, 참조 비디오가 없으면 입력 1초당 **단가의 2배**입니다. 참조 비디오가 있을 때도 실제 요금은 같은 식을 따릅니다. 라이브 배지는 비디오 길이를 알 수 없어 입력 1초당 **1배에서 2배** 범위를 표시합니다. 참조 이미지에는 장당 추가 요금이 없습니다.
+
+| 모델 | 해상도 | 단가(크레딧 / 초) | 참조 비디오 없음 | 참조 비디오 있음 |
+| :-------------- | :----- | ----------------: | :--------------- | :--------------- |
+| kling-4.0 | 720p | 28.36 | 56.73 / 입력 초(2 × 단가) | 28.36 × (입력 + 참조 초) |
+| kling-4.0 | 1080p | 42.54 | 85.09 / 입력 초(2 × 단가) | 42.54 × (입력 + 참조 초) |
+| kling-4.0-flash | 720p | 19.01 | 38.02 / 입력 초(2 × 단가) | 19.01 × (입력 + 참조 초) |
+
+참조 비디오가 없는 10초 1080p 편집은 850.9 크레딧입니다.
#### kling-video-o1
@@ -788,10 +873,6 @@ HeyGen Video 1.0 클립은 5~15초이며, 선택한 해상도의 요율로 청
| 모션 컨트롤 | — | std | 14.77 |
| 모션 컨트롤 | — | pro | 23.63 |
-
- **서비스 종료(EOL)**: Kling V1.5, V1.6, V2.1, V2.1 Master 및 Kolors 가상 착용은 **2026년 9월 15일**에 종료될 예정입니다. 아래 가격은 삭제될 때까지 참고용으로 유지됩니다.
-
-
#### kling-v2-5-turbo
**Kling 텍스트 투 비디오**, **Kling 이미지(첫 프레임) 투 비디오**, **Kling 시작-끝 프레임 투 비디오**는 `kling-v2-5-turbo` **pro** 모드만 사용하며, 고정 5초 또는 10초 길이로 실행당 청구됩니다.
@@ -839,6 +920,10 @@ HeyGen Video 1.0 클립은 5~15초이며, 선택한 해상도의 요율로 청
| :------------------- | :---------- |
| Kling Virtual Try-On | 33.79 |
+
+ **종료됨**: Kling은 **2026년 9월 15일**에 V1.5, V1.6, V2.1, V2.1 Master 및 이전 Kolors 가상 착용 서비스를 종료했습니다. Kling 노드는 더 이상 해당 모델을 제공하지 않으며, 이 페이지에서도 해당 가격 행을 제거했습니다. 위의 **Kling Virtual Try-On** 행은 별도의 현재 노드이며 실행당 청구됩니다.
+
+
## Lightricks
### 비디오: 크레딧 / 초
@@ -868,15 +953,35 @@ HeyGen Video 1.0 클립은 5~15초이며, 선택한 해상도의 요율로 청
해상도는 `1280x720` / `720x1280`(720p), `1920x1080` / `1080x1920`(1080p), `2560x1440` / `1440x2560`(1440p), `3840x2160` / `2160x3840`(4K)에 해당합니다. **LTX-2.5 (Fast)** 에서 10초를 초과하는 지속 시간은 720p 또는 1080p 해상도와 24/25 FPS가 필요합니다. **LTX-2 (Fast)** 에서 10초를 초과하는 지속 시간은 1080p 및 25 FPS로만 사용할 수 있습니다.
+### 후처리: 크레딧 / 초
+
+이 노드들은 입력 비디오를 처리하며 입력 길이(초)에 따라 과금됩니다.
+
+| 노드 | 해상도 | 크레딧 / 초 |
+| :------------------- | ------ | --------------- |
+| LTX Video AlphaGen | 1080p (짧은 변 ≤1080, 1080 미만은 이 요금) | 30.17 |
+| LTX Video AlphaGen | 1440p (짧은 변 ≤1440) | 75.43 |
+| LTX Video AlphaGen | 2160p (짧은 변 ≤2160) | 301.73 |
+| LTX Video SDR to HDR | 1080p (≤ 1920×1080) | 120.69 |
+| LTX Video SDR to HDR | 1440p (≤ 2560×1440) | 241.39 |
+| LTX Video SDR to HDR | 4K (≤ 3840×2160) | 482.77 |
+| LTX Video Reframe | 720p | 30.17 |
+| LTX Video Reframe | 1080p | 60.35 |
+| LTX Video Upscale | 1080p | 24.14 |
+| LTX Video Upscale | 4k | 90.52 |
+| LTX Video Upscale | 8k | 362.08 |
+
+**AlphaGen**은 입력 비디오의 초당 요금이며, 요금 구간은 입력의 짧은 변(너비와 높이 중 작은 쪽)으로 결정됩니다. 짧은 변이 1080픽셀 미만인 입력은 1080p 요금이 적용됩니다. **SDR to HDR**도 입력 비디오의 초당 요금이며, 입력의 픽셀 수를 담을 수 있는 가장 낮은 해상도 구간으로 맞춰집니다. 라이브 ComfyUI 배지는 입력 해상도를 볼 수 없어 구간만 표시합니다(AlphaGen은 초당 **30.17–301.73** 크레딧, SDR to HDR은 초당 **120.69–482.77** 크레딧). 실제 청구는 표의 3단계 요금을 따릅니다. **AlphaGen**은 최대 10초, **Reframe**은 최대 60초 비디오를 받고, **SDR to HDR**과 **Upscale**은 해상도별로 프레임 수를 제한합니다(SDR to HDR은 1080p에서 273프레임, Upscale은 2,401프레임).
+
## Luma
### 비디오
- **사용 중단**: 레거시 **Luma 텍스트 투 비디오** 및 **Luma 이미지 → 비디오** 노드는 사용이 중단되었습니다. Luma는 이 노드들이 사용하는 엔드포인트를 2026년 10월 24일에 제거했으므로 이 두 노드는 더 이상 비디오를 생성할 수 없습니다. 대신 **Luma Ray 3.2** 노드를 사용하세요. 아래의 ray-2, ray-flash-2, ray-1-6 행은 참고용으로만 유지됩니다. **ray-1-6**은 2026년 9월부터 제공되지 않습니다.
+ **지원 종료**: 레거시 Luma 텍스트 투 비디오, 이미지 → 비디오, 이미지 생성 노드는 제거되었고 ray-2, ray-flash-2, ray-1-6, photon-1, photon-flash-1 모델도 더 이상 제공되지 않습니다. 아래 표에는 ComfyUI가 현재 제공하는 노드만 정리했습니다.
-각 `duration` 선택에 대한 고정 **/실행** 가격은 노드 배지와 일치합니다. **편집** 및 **리프레임**은 **(크레딧 / 초) × 비디오 길이**로 청구됩니다. **ray-1-6**은 항상 **105.50 / 실행**입니다.
+각 `duration` 선택에 대한 고정 **/실행** 가격은 노드 배지와 일치합니다. **편집** 및 **리프레임**은 **(크레딧 / 초) × 비디오 길이**로 청구됩니다.
| 모델 | 작업 | 해상도 | 크레딧 |
| :---------- | :--------- | :--------- | :------------------------------------------------ |
@@ -892,24 +997,13 @@ HeyGen Video 1.0 클립은 5~15초이며, 선택한 해상도의 요율로 청
| ray-3.2 | 리프레임 | 540p | 12.66 / 초 |
| ray-3.2 | 리프레임 | 720p | 25.32 / 초 |
| ray-3.2 | 리프레임 | 1080p | 75.96 / 초 |
-| ray-2 | 생성 | 540p | 120.27 / 실행 (5초), 217.33 / 실행 (9초) |
-| ray-2 | 생성 | 720p | 215.22 / 실행 (5초), 386.13 / 실행 (9초) |
-| ray-2 | 생성 | 1080p | 478.97 / 실행 (5초), 865.10 / 실행 (9초) |
-| ray-2 | 생성 | 4K | 1922.21 / 실행 (5초), 3460.40 / 실행 (9초) |
-| ray-flash-2 | 생성 | 540p | 42.20 / 실행 (5초), 75.96 / 실행 (9초) |
-| ray-flash-2 | 생성 | 720p | 71.74 / 실행 (5초), 128.71 / 실행 (9초) |
-| ray-flash-2 | 생성 | 1080p | 166.69 / 실행 (5초), 299.62 / 실행 (9초) |
-| ray-flash-2 | 생성 | 4K | 660.43 / 실행 (5초), 1192.15 / 실행 (9초) |
-| ray-1-6 | 생성 | — | 105.50 / 실행 |
-
-**Luma Ray 3.2** 이미지를 비디오로 및 비디오 확장은 항상 **5초** 가격으로 청구됩니다. 레거시 이미지를 비디오로는 **ray-2** / **ray-flash-2** 행의 **5초** 또는 **9초** 가격을 사용합니다.
+
+**Luma Ray 3.2** 이미지를 비디오로 및 비디오 확장은 항상 **5초** 가격으로 청구됩니다.
### 이미지
| 모델 | 작업 | 크레딧 |
| :------------- | :--------- | :----------------------------------- |
-| photon-flash-1 | — | 0.57 / 실행 |
-| photon-1 | — | 2.19 / 실행 |
| uni-1 | 생성 | 8.52 / 실행 + 0.63 / 참조 이미지 (≤ 9) |
| uni-1 | 편집 | 9.16 / 실행 + 0.63 / 참조 이미지 (≤ 8) |
| uni-1-max | 생성 | 21.10 / 실행 + 0.63 / 참조 이미지 (≤ 9) |
@@ -975,7 +1069,7 @@ Ultra 모드(Meshy 7.1, Meshy 7, `latest` 모델에서 사용 가능)는 60.3
| MiniMax H3 첫·끝 프레임 투 비디오 | MiniMax H3 | 2K | 39.22 / 초 |
| MiniMax H3 레퍼런스 투 비디오 | MiniMax H3 | 768P | 27.16 / 초, 5장 초과 시 장당 +12.07 |
| MiniMax H3 레퍼런스 투 비디오 | MiniMax H3 | 2K | 39.22 / 초, 5장 초과 시 장당 +12.07 |
-| MiniMax H3 컨텍스트 IR | MiniMax H3 | — | 4.22 / 실행(텍스트만), 10.55 / 실행(첫·끝 프레임), 12.66–23.21 / 실행(참조 미디어) |
+| MiniMax H3 컨텍스트 IR | MiniMax H3 | — | 271.56 / 입력 토큰 100만 + 1,086.23 / 출력 토큰 100만 |
| MiniMax H3 2K 재생성 | MiniMax H3 | — | 15.09 / 초 |
| MiniMax H3 텍스트 투 비디오 | MiniMax H3 Max | 480P | 15.09 / 초 |
| MiniMax H3 텍스트 투 비디오 | MiniMax H3 Max | 768P | 24.14 / 초 |
@@ -1008,7 +1102,7 @@ Ultra 모드(Meshy 7.1, Meshy 7, `latest` 모델에서 사용 가능)는 60.3
**MiniMax Hailuo 02 Video**는 1080P에서 **6초**만 지원합니다.
-**MiniMax H3**는 초 단위로 청구됩니다(4~15초). **MiniMax H3 레퍼런스 투 비디오**는 참조 이미지 5장 초과 시 장당 12.07크레딧, 참조 비디오는 27.16크레딧/초(768P) 또는 39.22크레딧/초(2K)로 청구됩니다(비디오당 2초부터, 총 최대 15초, 근사값). **MiniMax H3 컨텍스트 IR** 요금은 첨부 미디어(텍스트만, 첫·끝 프레임, 참조 미디어)에 따라 달라집니다(근사값).
+**MiniMax H3**는 초 단위로 청구됩니다(4~15초). **MiniMax H3 레퍼런스 투 비디오**는 참조 이미지 5장 초과 시 장당 12.07크레딧, 참조 비디오는 27.16크레딧/초(768P) 또는 39.22크레딧/초(2K)로 청구됩니다(비디오당 2초부터, 총 최대 15초, 근사값). **MiniMax H3 컨텍스트 IR**은 **입력 토큰 100만당 271.56 크레딧**, **출력 토큰 100만당 1,086.23 크레딧**으로 청구됩니다. 노드 배지는 그래프를 구성할 때 토큰 수를 알 수 없으므로, 대략 **4.22**(텍스트만), **10.55**(첫·끝 프레임), **12.66–23.21**(참조 미디어)의 범위를 표시합니다.
**MiniMax H3 Max**는 위 표의 초당 요율로 청구됩니다. **텍스트 투 비디오**, **첫·끝 프레임 투 비디오**, **레퍼런스 투 비디오** 출력은 **5~15초**입니다. **립싱크**는 5~약 899초 오디오를 지원하며, 출력이 15초를 넘으면 아래 **1.2×** 배율이 적용됩니다.
@@ -1034,6 +1128,19 @@ Ultra 모드(Meshy 7.1, Meshy 7, `latest` 모델에서 사용 가능)는 60.3
**MiniMax H3 Max 인물 교체**, **비디오 연장**, **리프레임**, **비디오 삽입**, **립싱크**, **카메라 제어**는 비디오 초 단위로 과금됩니다. 비디오 연장은 Turbo 모델도 선택할 수 있으며 요금은 Max의 절반입니다. **리프레임**은 표의 요율로 초당 과금됩니다. 긴 소스는 겹치는 구간을 사용하므로 과금 초가 최종 출력 길이를 넘을 수 있습니다. **립싱크**는 표의 요율로 초당 과금됩니다. 출력이 **15초를 초과**하면 **1.2×** 배율이 적용됩니다(예: 768P → 28.97 / 초). 비디오 삽입은 무료 4,096토큰 한도를 넘는 참조 토큰에 1K 토큰당 6.03 크레딧이 붙으며, 레퍼런스 투 비디오와 같은 한도를 공유합니다(기본 240토큰, 이미지당 2,048토큰, 참조 비디오는 768P 출력에서 225,000토큰, 480P에서 87,000토큰).
+### 이미지: MiniMax H3 Image
+
+생성된 이미지당 가격입니다. 출력 크기에 더해 프롬프트와 참조 이미지가 토큰으로 과금됩니다.
+
+| 노드 | 모델 | 크기 | 크레딧 / 이미지 |
+| :--------------- | -------------------------- | ---- | --------------- |
+| MiniMax H3 Image | MiniMax-H3-Image | 2K | 5.32 – 8.17 |
+| MiniMax H3 Image | MiniMax-H3-Image | 4K | 19.41 – 22.26 |
+| MiniMax H3 Image | MiniMax-H3-Image-Highspeed | 2K | 10.02 – 12.86 |
+| MiniMax H3 Image | MiniMax-H3-Image-Highspeed | 4K | 38.20 – 41.04 |
+
+위 표는 텍스트 전용이며 `enhance_prompt`가 켜진 경우의 요금입니다. 전체 계산식은 다음과 같습니다. 출력 크레딧 = `max(픽셀 수, 4,194,304) / 1,024 × 0.00114657`(Highspeed는 `0.00229315`). 2K(4 MP)는 **4.70 크레딧**, 4K는 약 **18.79 크레딧**, Highspeed는 **9.39** / **37.57**입니다. 프롬프트 토큰은 토큰당 **0.000694 크레딧**이며, 참조 이미지 없이 확장이 꺼지면 **0–300**, 참조 이미지 없이 확장이 켜지면 **900–5,000**, 참조 이미지가 `n`장일 때는 **2,000 + 300 × (n − 1)**에서 **6,500 + 800 × (n − 1)**입니다. 각 참조 이미지는 여기에 **2,050–8,194**개의 이미지 토큰을 토큰당 **0.000453 크레딧**(Highspeed는 **0.000905**)으로 더하며, 약 **0.93~3.71 크레딧**(Highspeed는 **1.86~7.42 크레딧**)입니다. 확장을 끈 2K 이미지는 약 **4.70 크레딧**입니다. 사용자 지정 크기는 해당 면적으로 계산되며 최소 4메가픽셀이 적용됩니다.
+
## OpenAI
DALL·E와 Sora는 이미지당 또는 초당 고정 요금을 사용합니다. **OpenAI GPT Image 2** 및 **OpenAI ChatGPT**는 API 토큰 사용량에 따라 청구됩니다.
@@ -1084,14 +1191,55 @@ DALL·E와 Sora는 이미지당 또는 초당 고정 요금을 사용합니다.
출력 토큰 수는 크기와 품질에 따라 달라집니다(예: 1024x1024 `low`는 약 196개의 출력 토큰이며, `gpt-image-1.5` 기준 약 1.32 크레딧). 1024x1024 참조 이미지 1장당 약 2.47 크레딧의 이미지 입력이 추가되며, 프롬프트 텍스트는 텍스트 입력 요금으로 청구됩니다. 생성 수량(`n`)에 따라 총액이 비례합니다.
+**GPT Image 2.6** 모델(`gpt-image-2.6-flare`, `gpt-image-2.6-sunburst`)은 동일한 요금표를 사용합니다. 아래는 1024x1024 이미지 1장당 크레딧입니다.
+
+| 품질 | 크레딧 / 장 |
+| :----- | :---------- |
+| low | 7.45 |
+| medium | 13.12 |
+| high | 15.89 |
+| xhigh | 28.25 |
+| max | 63.57 |
+
+크기가 클수록 비쌉니다. 3840x2160은 `low`에서 14.12 크레딧, `max`에서 120.8 크레딧입니다. 총 크레딧 = **(출력 가격 + 참조 이미지 수 × 2.47~3.71) × `n`**. 참조 이미지 1장당 **2.47~3.71 크레딧**이 추가됩니다. 라이브 배지는 해당 단가를 결정하는 참조 이미지 크기를 볼 수 없어 이 구간을 표시합니다. 생성 수량(`n`)에 따라 총액이 비례합니다.
+
+### 이미지: GPT Image 2.6 GIF
+
+**OpenAI GPT Image 2.6 GIF**는 토큰이 아니라 실행당 청구됩니다. 프레임마다 이미지 한 장이므로 프레임 크기, 품질, `frame_count`가 가격을 결정합니다. `frame_count`는 2~8프레임, 9프레임, 10~12프레임, 13~16프레임 네 구간으로 청구되며, 큰 사이즈일수록 선택할 수 있는 구간이 적습니다. 512x512, 9프레임 기준 크레딧:
+
+| 품질 | 크레딧 / 실행 |
+| :----- | :------------ |
+| low | 10.61 |
+| medium | 18.76 |
+| high | 22.72 |
+| xhigh | 40.41 |
+| max | 90.92 |
+
+384x384에서 2~8프레임은 `low`에서 3.88 크레딧, 720x720에서 13~16프레임은 `max`에서 214.69 크레딧입니다. `frame_duration_ms`는 가격에 영향을 주지 않습니다.
+
+### 이미지: GPT Image 2.6 SVG
+
+**OpenAI GPT Image 2.6 SVG**는 선택한 크기와 품질의 정지 이미지 가격에 벡터 처리분 **0.23~1.77 크레딧**을 더한 금액입니다. 라이브 배지는 벡터 처리의 정확한 요금을 결정하는 입력을 볼 수 없어, 정지 이미지 가격 위에 이 구간을 표시합니다. 표의 범위는 정지 이미지 가격에 그 구간을 더한 값입니다. SVG 배지에는 별도의 참조 이미지 요금이 포함되지 않습니다. 사이즈는 1024x1024, 1024x1536, 1536x1024입니다. 1024x1024 기준 크레딧:
+
+| 품질 | 크레딧 / 실행 |
+| :----- | :------------ |
+| low | 7.68 – 9.22 |
+| medium | 13.36 – 14.90 |
+| high | 16.12 – 17.66 |
+| xhigh | 28.48 – 30.03 |
+| max | 63.81 – 65.35 |
+
+1024x1536 또는 1536x1024는 더 저렴합니다. `low`에서 6.25~7.79 크레딧, `max`에서 49.90~51.44 크레딧입니다.
+
### 채팅
토큰 기반 청구입니다. 배지 요금은 **1K 토큰당**이며, 표는 **1M 토큰당**입니다.
| 모델 | 입력 크레딧 / 1M | 출력 크레딧 / 1M |
| :----------- | -----------------: | :------------------ |
+| gpt-6.1-sol、gpt-6-sol | 603.46 | 3017.3 |
+| gpt-6.1-terra | 301.73 | 1508.65 |
| gpt-6-astra | 3017.3 | 15086.5 |
-| gpt-6-sol | 603.46 | 3017.3 |
| gpt-6-luna | 30.17 | 150.87 |
| gpt-5.6-sol、gpt-5.6-sol-pro | 1055 | 6330 |
| gpt-5.6-terra、gpt-5.6-terra-pro | 527.5 | 3165 |
@@ -1151,10 +1299,12 @@ DALL·E와 Sora는 이미지당 또는 초당 고정 요금을 사용합니다.
| x-ai/grok-4.5 | 603.46 | 1810.38 |
| x-ai/grok-4.20 | 377.16 | 754.33 |
| x-ai/grok-4.3 | 377.16 | 754.33 |
+| meta/muse-wonder | 1206.92 | 6034.60 |
+| meta/muse-spark-1.3 | 377.16 | 1282.35 |
+| meta/muse-ember | 30.17 | 150.87 |
| deepseek/deepseek-v4-pro | 131.25 | 262.51 |
| deepseek/deepseek-v4-flash | 33.79 | 67.59 |
| deepseek/deepseek-v3.2 | 76.04 | 114.05 |
-| qwen/qwen3.6-max-preview | 313.80 | 1882.80 |
| qwen/qwen3.6-plus | 98.06 | 588.37 |
| qwen/qwen3.6-flash | 56.57 | 339.45 |
| mistralai/mistral-large-2512 | 150.87 | 452.60 |
@@ -1196,6 +1346,39 @@ DALL·E와 Sora는 이미지당 또는 초당 고정 요금을 사용합니다.
가격은 근사치입니다. 1 MP 머티리얼에 기본 `basecolor`, `normal` 맵만 켜면 15.09 크레딧이고, 4 MP에서 다섯 맵을 모두 4× 업스케일하면 **Patina Material**은 184.06 크레딧, **Patina Material Extract**는 211.21 크레딧입니다.
+## Pika
+
+Pika H3 Flex는 비디오 초 단위로 청구됩니다. 총 크레딧 = **(크레딧 / 초) × 청구되는 초**. `duration`은 5~15초입니다. `480p`와 `768p`에서는 청구 길이가 `duration`과 같습니다. `1080p` 이상에서는 노드가 출력을 프레임 그리드에 맞추므로 5초 생성은 5.17초, 10초는 10.13초, 15초는 15.08초로 청구됩니다.
+
+| 노드 | 해상도 | 크레딧 / 초 |
+| :------------------------------ | :----- | :---------- |
+| Pika H3 Flex Video | 480p | 3.02 |
+| Pika H3 Flex Video | 768p | 6.03 |
+| Pika H3 Flex Video | 1080p | 10.56 |
+| Pika H3 Flex Video | 2k | 12.07 |
+| Pika H3 Flex Video | 4k | 24.14 |
+| Pika H3 Flex Reference to Video | 480p | 9.05 |
+| Pika H3 Flex Reference to Video | 768p | 18.10 |
+| Pika H3 Flex Reference to Video | 1080p | 22.63 |
+| Pika H3 Flex Reference to Video | 2k | 24.14 |
+| Pika H3 Flex Reference to Video | 4k | 30.17 |
+
+**Pika H3 Flex Video**를 480p로 5초 생성하면 15.09 크레딧, 1080p로 5초 생성하면 54.56 크레딧입니다.
+
+**Pika H3 Flex Reference to Video**는 연결한 참조 비디오도 같은 요율로 청구합니다. 각 클립은 2~15초이고 참조 비디오 총 길이는 15초까지입니다. 따라서 요금은 연결한 참조 비디오의 길이에 따라 달라집니다. 480p로 5초 생성하고 2초 참조 비디오 1개를 연결하면 63.36 크레딧입니다. 4k로 5초 생성하고 참조 비디오를 총 15초 연결하면 608.49 크레딧입니다. 위 Pika 예시는 모두 반올림하지 않은 초당 요율로 계산했습니다.
+
+### 오디오: Pika
+
+Pika 오디오 노드도 초 단위로 청구됩니다. **Pika Soundtrack**은 입력 비디오 길이(최대 10분), **Pika SFX**는 `duration`(1~20초), **Pika Music**은 곡 길이로 청구됩니다. `text to music`은 `duration`(10~360초)을, `cover`는 원본 오디오 길이(5~360초로 제한)를 사용합니다.
+
+| 노드 | 크레딧 / 초 |
+| :-------------- | :---------- |
+| Pika Soundtrack | 1.509 |
+| Pika SFX | 0.060 |
+| Pika Music | 0.075 |
+
+10분 Soundtrack은 905.2 크레딧, 20초 SFX는 1.21 크레딧, 3분 Music은 13.58 크레딧입니다.
+
## Pixverse
모델 **v3.5** (API에서 고정됨, 선택 불가).
@@ -1220,19 +1403,12 @@ PixVerse V6 노드는 선택적 네이티브 오디오와 함께 1~15초 길이
| PixVerse V6 | 540p | 7.18 | 9.23 |
| PixVerse V6 | 720p | 9.23 | 12.31 |
| PixVerse V6 | 1080p | 18.47 | 23.60 |
-| PixVerse V6 Max | 360p | 6.71 | 9.39 |
-| PixVerse V6 Max | 540p | 9.39 | 12.07 |
-| PixVerse V6 Max | 720p | 12.07 | 16.09 |
-| PixVerse V6 Max | 1080p | 24.14 | 30.84 |
-
-**PixVerse V6 Max**는 **2026년 11월 3일** 전까지 위 요금이 도입 가격으로 적용됩니다. 같은 날부터의 초당 요금은 다음과 같습니다:
+| PixVerse V6 Max | 360p | 8.05 | 12.07 |
+| PixVerse V6 Max | 540p | 12.07 | 14.75 |
+| PixVerse V6 Max | 720p | 14.75 | 20.12 |
+| PixVerse V6 Max | 1080p | 29.50 | 37.55 |
-| 화질 | 오디오 없음 | 오디오 있음 |
-| :---- | :---------- | :---------- |
-| 360p | 8.05 | 12.07 |
-| 540p | 12.07 | 14.75 |
-| 720p | 14.75 | 20.12 |
-| 1080p | 29.50 | 37.55 |
+**PixVerse V6 Max**의 도입 가격은 **2026년 11월 2일**까지 적용되었습니다. 위 표는 **2026년 11월 3일**부터 적용되는 표준 요금입니다.
**PixVerse V6 Fusion**은 출력 길이가 아닌 참조 비디오의 길이를 기준으로 초당 요금의 2배를 청구합니다. Fusion은 `PixVerse V6` 요금만 사용합니다.
@@ -1240,21 +1416,10 @@ PixVerse V6 노드는 선택적 네이티브 오디오와 함께 1~15초 길이
### 이미지
-종료된 **arrow-1.1** 모델은 실행당 고정 요금이 부과되었습니다:
-
- **더 이상 사용되지 않음**: Quiver는 **2026년 10월 16일**에 Arrow 1 시리즈 모델을 종료했으며, Text to SVG와 Image to SVG 노드에서 `arrow-1.1`, `arrow-1.1-max`, `arrow-preview`를 더 이상 선택할 수 없습니다. 대신 Arrow 2 모델을 사용하세요. 아래 요금은 참고용으로만 남겨 둡니다.
+ **종료됨**: Quiver는 **2026년 10월 16일**에 Arrow 1 시리즈 모델(`arrow-1.1`, `arrow-1.1-max`, `arrow-preview`)을 종료했으며, Text to SVG와 Image to SVG 노드에서 더 이상 선택할 수 없습니다. 아래 Arrow 2 모델을 사용하세요.
-| 모델 | 노드 | 크레딧 / 실행 |
-| :------------ | :------------------ | :---------------- |
-| arrow-1.1 | Quiver Text to SVG | 60.35 |
-| arrow-1.1 | Quiver Image to SVG | 45.26 |
-| arrow-1.1-max | Quiver Text to SVG | 75.43 |
-| arrow-1.1-max | Quiver Image to SVG | 60.35 |
-| arrow-preview | Quiver Text to SVG | 90.52 |
-| arrow-preview | Quiver Image to SVG | 90.52 |
-
**arrow-2**와 **arrow-2-telos**는 요청이 소비한 토큰을 기준으로 청구됩니다:
| 모델 | 입력 크레딧 / 1M 토큰 | 출력 크레딧 / 1M 토큰 |
@@ -1375,12 +1540,15 @@ Gen4 Turbo: **5초** 또는 **10초**. 총 크레딧 = **(크레딧 / 초) × `d
## Sonilo
-Sonilo 텍스트로 음악 생성: 총 크레딧 = **(크레딧 / 초) × `duration`** (1–360초).
+비디오로 음악 생성은 입력 비디오 초 단위로 청구되며 최대 6분입니다. 텍스트로 음악 생성과 Music 1.5는 오디오 초당 청구됩니다: 총 크레딧 = **(크레딧 / 초) × `duration`**.
| 노드 | 크레딧 |
| :-------------------- | :----------- |
-| Sonilo 비디오로 음악 생성 | 1.9 / 초 |
-| Sonilo 텍스트로 음악 생성 | 0.5275 / 초 |
+| Sonilo 비디오로 음악 생성 | 2.72 / 초 |
+| Sonilo 텍스트로 음악 생성 | 0.6789 / 초 |
+| Sonilo Music 1.5 | 0.6789 / 초 |
+
+텍스트로 음악 생성의 길이는 1~360초입니다. Music 1.5는 `auto`로 가사에 맞는 길이를 정하거나 `custom`으로 10~360초를 지정할 수 있습니다.
## sync.so
@@ -1451,7 +1619,7 @@ Multiview는 `image_left`, `image_right` 또는 `image_back`이 연결된 경우
| :-------- | :------------- |
| Reimagine | 67.52 / 8MP |
| Bloom 2 | 96.55 / 8MP |
-| Wonder 3.5 | 24.14 / 8MP |
+| Wonder 3.5, Wonder 4 | 24.14 / 8MP |
**Topaz Image Enhance(레거시)**는 더 이상 사용되지 않습니다. 실행 전 고정 견적은 없으며, 폴링 시점에 API가 반환하는 견적(API 크레딧 × 0.08 USD)에 따라 청구됩니다. 입력 해상도별 일반적인 요금:
@@ -1612,6 +1780,19 @@ Pruna P-Video-2는 생성된 비디오의 초당 청구됩니다. 총 크레딧
| | 1080p | 꺼짐 | 15.09 |
| | 1080p | 켜짐 | 9.05 |
+### 비디오: P-Video Upscale
+
+Pruna P-Video Upscale는 출력 비디오의 초당 청구됩니다. 총 크레딧 = **(크레딧 / 초) × 출력 길이**. 요금은 `target`에 따라 달라지며, 30 fps를 초과하는 출력은 두 배의 요금이 적용됩니다. 원본이 대상보다 크면 원래 크기로 출력됩니다.
+
+`720p`, `1080p`, `2K`, `4K`에는 2026년 11월 12일(17:00 UTC)까지 초당 1.51, 3.02, 6.03, 12.07 크레딧의 출시 요금이 적용되었습니다. 아래 표는 표준 요금이며 같은 날부터 적용됩니다.
+
+| target | 대상 크기 | 크레딧 / 초 |
+| :----- | :-------- | :---------- |
+| 720p | 약 0.9 MP | 3.02 |
+| 1080p | 약 2.1 MP | 6.03 |
+| 2K | 약 3.7 MP | 12.07 |
+| 4K | 약 8.3 MP | 24.14 |
+
## VEED
| 노드 | 옵션 | 크레딧 |
@@ -1701,23 +1882,25 @@ Pruna P-Video-2는 생성된 비디오의 초당 청구됩니다. 총 크레딧
| wan2.5-t2i-preview, wan2.5-i2i-preview | — | 6.33 / 실행 |
| wan2.7-t2v, wan2.7-i2v | 720P | 21.1 / 초 |
| | 1080P | 31.65 / 초 |
-| wan2.7-i2v (continuation) | 720P | 21.1 / 초 (범위) |
-| | 1080P | 31.65 / 초 (범위) |
+| wan2.7-i2v (continuation) | 720P | 21.1 / 초 (출력 + 입력 2–5초) |
+| | 1080P | 31.65 / 초 (출력 + 입력 2–5초) |
| wan2.7-videoedit | 720P | 21.1 / 초 (입력 + 출력) |
| | 1080P | 31.65 / 초 (입력 + 출력) |
-| wan2.7-r2v | 720P | 21.1 / 초 (범위) |
-| | 1080P | 31.65 / 초 (범위) |
-| wan3.0-video | 480P | 15.09 / 초 (범위) |
-| | 720P | 30.17 / 초 (범위) |
-| | 1080P | 60.35 / 초 (범위) |
-| wan3.0-video-prime | 480P | 20.52 / 초 (범위) |
-| | 720P | 42.24 / 초 (범위) |
-| | 1080P | 84.48 / 초 (범위) |
+| wan2.7-r2v | 720P | 21.1 / 초 (출력 ~ 출력 + 입력 5초) |
+| | 1080P | 31.65 / 초 (출력 ~ 출력 + 입력 5초) |
+| wan3.0-video | 480P | 15.09 / 초 |
+| | 720P | 30.17 / 초 |
+| | 1080P | 60.35 / 초 |
+| wan3.0-video-prime | 480P | 20.52 / 초 |
+| | 720P | 42.24 / 초 |
+| | 1080P | 84.48 / 초 |
| wan2.5-t2v-preview, wan2.5-i2v-preview | 480p | 10.55 / 초 |
| wan2.5-t2v-preview, wan2.5-i2v-preview, wan2.6-t2v, wan2.6-i2v | 720p | 21.1 / 초 |
| wan2.5-t2v-preview, wan2.5-i2v-preview, wan2.6-t2v, wan2.6-i2v | 1080p | 31.65 / 초 |
-| wan2.6-r2v | 720p | 21.1 / 초 (범위) |
-| | 1080p | 31.65 / 초 (범위) |
+| wan2.6-r2v | 720p | 21.1 / 초 (출력 + 입력 2–5초) |
+| | 1080p | 31.65 / 초 (출력 + 입력 2–5초) |
+
+**Wan 2.6 Reference to Video**와 **Wan 2.7 continuation**은 출력 길이에 입력 **2–5초**를 더한 시간을 표와 같은 초당 요금으로 청구합니다. **Wan 2.7 Reference to Video**는 출력 길이만부터 출력 길이에 입력 **5초**를 더한 범위입니다. **Wan 2.7 Edit**는 입력과 출력을 표의 요금으로 청구합니다. **Wan 3.0**은 표의 초당 요금으로 청구되며, `duration`이 `auto`이면 배지에 그 초당 요금이 표시되고, 그렇지 않으면 선택한 길이부터 `min(duration + 15, 30)`초까지의 범위입니다.
## WaveSpeed
@@ -1742,17 +1925,23 @@ Pruna P-Video-2는 생성된 비디오의 초당 청구됩니다. 총 크레딧
| :------------------------- | :--------- | :------- | :------------------------------------------- |
| grok-imagine-image-2.0 | 1K | 생성 | 12.66(중간)/ 8.44(낮음) 출력 이미지당 |
| | 2K | 생성 | 16.88(중간)/ 12.66(낮음) 출력 이미지당 |
-| grok-imagine-image-quality | 1K | 생성 | 10.55 / 출력 이미지 |
-| | 2K | 생성 | 14.77 / 출력 이미지 |
-| grok-imagine-image-pro | — | 생성 | 14.77 / 출력 이미지 |
+| grok-imagine-image-quality | 1K | 생성 | 8.44 / 출력 이미지 |
+| | 2K | 생성 | 12.66 / 출력 이미지 |
+| grok-imagine-image-pro | 1K | 생성 | 8.44 / 출력 이미지 |
+| | 2K | 생성 | 12.66 / 출력 이미지 |
| grok-imagine-image | — | 생성 | 4.22 / 출력 이미지 |
-| grok-imagine-image-2.0 | 1K | 편집 | 2.11 입력 + 12.66(중간)/ 8.44(낮음) 출력 이미지당 |
-| | 2K | 편집 | 2.11 입력 + 16.88(중간)/ 12.66(낮음) 출력 이미지당 |
-| grok-imagine-image-quality | 1K | 편집 | 2.11 입력 + 10.55 / 출력 이미지 |
-| | 2K | 편집 | 2.11 입력 + 14.77 / 출력 이미지 |
-| grok-imagine-image-pro | — | 편집 | 2.11 입력 + 14.77 / 출력 이미지 |
+| grok-imagine-image-2.0 | 1K | 편집 | 2.11–6.33 입력 + 12.66(중간)/ 8.44(낮음) 출력 이미지당 |
+| | 2K | 편집 | 2.11–6.33 입력 + 16.88(중간)/ 12.66(낮음) 출력 이미지당 |
+| grok-imagine-image-quality | 1K | 편집 | 2.11–6.33 입력 + 8.44 / 출력 이미지 |
+| | 2K | 편집 | 2.11–6.33 입력 + 12.66 / 출력 이미지 |
+| grok-imagine-image-pro | 1K | 편집 | 2.11 입력 + 8.44 / 출력 이미지 |
+| | 2K | 편집 | 2.11 입력 + 12.66 / 출력 이미지 |
| grok-imagine-image | — | 편집 | 0.422–1.266 입력 + 4.22 / 출력 이미지 |
+프로가 아닌 편집 모델은 입력 이미지당 **2.11 크레딧**이며, 입력 이미지는 **1~3장**입니다(**2.11–6.33**). `grok-imagine-image`는 입력 이미지당 **0.422 크레딧**입니다(**0.422–1.266**). 프로 편집 행은 입력 이미지 1장만 **2.11 크레딧**으로 청구합니다.
+
+`grok-imagine-image-quality`와 `grok-imagine-image-pro`는 xAI에서 서비스가 종료되었습니다. 이미지 노드는 두 모델을 계속 표시하고 `grok-imagine-image-2.0`의 낮음 품질 요율로 청구합니다.
+
### 비디오
| 모델 | 해상도 | 크레딧 |
@@ -1770,6 +1959,8 @@ Pruna P-Video-2는 생성된 비디오의 초당 청구됩니다. 총 크레딧
| | 720p | 21.12 / 초 출력 + 0.603 / 참조 이미지 |
| grok-imagine-video (extend)| — | 6.03–45.26 입력 + 15.09 × 확장 초 출력 / 실행 |
+**Extend**는 확장 초당 **15.09 크레딧**에 더해, 소스 영상 입력 요금으로 2초일 때 **6.03 크레딧**, 15초일 때 **45.26 크레딧**을 청구합니다(소스 영상은 2–15초여야 합니다). 라이브 배지는 소스 길이를 볼 수 없어, 이 **6.03–45.26** 입력 구간에 초당 확장 요금을 더해 표시합니다.
+
## 클라우드 GPU
| GPU 유형 | 크레딧 |
diff --git a/tutorials/partner-nodes/pricing.mdx b/tutorials/partner-nodes/pricing.mdx
index 75ebdd40c8..e83bbc914c 100644
--- a/tutorials/partner-nodes/pricing.mdx
+++ b/tutorials/partner-nodes/pricing.mdx
@@ -30,6 +30,7 @@ Partner Nodes are billed per call in credits. Comfy API GPU time and Comfy Cloud
| :--------- | -----------------: | --------------------------: | --------------------------: | ---------------------: | :------------------ |
| Opus 5.5 | 1.207 | 1.509 | 2.414 | 0.121 | 6.035 |
| Opus 5, Opus 4.8 | 1.509 | 1.886 | 3.017 | 0.151 | 7.543 |
+| Fable 5.2 | 2.414 | 3.017 | 4.828 | 0.241 | 12.069 |
| Fable 5.1, Fable 5 | 3.017 | 3.772 | 6.035 | 0.302 | 15.087 |
| Sonnet 5.5, Sonnet 5 | 0.603 | 0.754 | 1.207 | 0.060 | 3.017 |
@@ -55,17 +56,19 @@ Optional images add tokens billed at the same model input rates.
### Image
-| Node | Credits |
-| :------------------------- | :---------------------------: |
-| Flux 1.1 [pro] Ultra Image | 12.66 / run |
-| Flux.1 Kontext [pro] Image | 8.44 / run |
-| Flux.1 Kontext [max] Image | 16.88 / run |
-| Flux.1 Expand Image | 10.55 / run |
-| Flux.1 Fill Image | 10.55 / run |
-| Flux Erase Image | 6.33 / run + 0.84 / extra MP |
-| Flux Virtual Try-On | 7.91 / run + 1.06 / extra MP |
-| FLUX.2 [pro] | 6.33 / run + 3.17 / extra MP |
-| FLUX.2 [max] | 14.77 / run + 6.33 / extra MP |
+| Node | Credits | Credits, with reference images |
+| :------------------------- | :---------------------------: | :-----------------------------------: |
+| Flux 1.1 [pro] Ultra Image | 12.66 / run | — |
+| Flux.1 Kontext [pro] Image | 8.44 / run | — |
+| Flux.1 Kontext [max] Image | 16.88 / run | — |
+| Flux.1 Expand Image | 10.55 / run | — |
+| Flux.1 Fill Image | 10.55 / run | — |
+| Flux Erase Image | 6.33 / run + 0.84 / extra MP | — |
+| Flux Virtual Try-On | 7.91 / run + 1.06 / extra MP | — |
+| FLUX.2 [pro] | 6.33 / run + 3.17 / extra MP | 9.50 – 31.65 / run + 3.17 / extra MP |
+| FLUX.2 [max] | 14.77 / run + 6.33 / extra MP | 21.10 – 65.41 / run + 6.33 / extra MP |
+
+Connecting reference images moves **FLUX.2** to the third column. Text-only **FLUX.2 [pro]** is **6.33 credits** plus **3.17 credits** per extra output megapixel (`ceil(W × H / 1 MP) − 1`). **FLUX.2 [max]** is **14.77 credits** plus **6.33 credits** per extra output megapixel. With reference images, add **3.17 credits** on [pro] or **6.33 credits** on [max] per input megapixel, rounded up: sum `ceil(W × H / 1 MP)` across up to 8 references. The third column is the 1 MP output envelope for 1–8 megapixels of reference input (**9.50–31.65** / **21.10–65.41**), plus the same per-extra-output-megapixel charge as the text-only column. The **3.17** unit on [pro] is rounded; the published **9.50–31.65** envelope is the live badge endpoint for 1–8 MP of reference input.
### Image: FLUX 3 Image
@@ -100,14 +103,14 @@ Total credits = **(credits / sec) × `duration`** (`duration` is 5–20 seconds
### Video: FLUX 3 Max
-Total credits = **(credits / sec) × `duration`** (`duration` is 5–20 seconds). `reasoning_level` and `generate_audio` do not change the price. Launch promo rates apply until **November 7, 2026 (00:00 UTC)**; standard rates apply after that.
+Total credits = **(credits / sec) × `duration`** (`duration` is 5–20 seconds). `reasoning_level` and `generate_audio` do not change the price. A launch rate of 12.82 credits per second at 720p and 21.88 at 1080p applied through November 7, 2026 (00:00 UTC). The rows below list the standard rates, which apply from that date.
-| Node | resolution | Promo credits / sec | Standard credits / sec |
-| :------------------------ | :--------- | :------------------ | :--------------------- |
-| FLUX 3 Max Text to Video | 720p | 12.82 | 51.29 |
-| FLUX 3 Max Text to Video | 1080p | 21.88 | 87.50 |
-| FLUX 3 Max Image to Video | 720p | 12.82 | 51.29 |
-| FLUX 3 Max Image to Video | 1080p | 21.88 | 87.50 |
+| Node | resolution | Credits / sec |
+| :------------------------ | :--------- | :------------ |
+| FLUX 3 Max Text to Video | 720p | 51.29 |
+| FLUX 3 Max Text to Video | 1080p | 87.50 |
+| FLUX 3 Max Image to Video | 720p | 51.29 |
+| FLUX 3 Max Image to Video | 1080p | 87.50 |
### Video: Flux Video Edit
@@ -123,10 +126,10 @@ Rates are **per output megapixel per second**. Total credits = **credits / MP /
| Node | mode | Credits / MP / sec |
| :----------------- | :-------- | :----------------- |
-| Flux Video Upscale | creative | 15.09 |
-| Flux Video Upscale | precise | 10.56 |
+| Flux Video Upscale | creative | 14.39 |
+| Flux Video Upscale | precise | 10.07 |
-A 1080p output (2.07 MP) costs 31.24 credits per second in `creative` mode and 21.86 in `precise`. A 4K output (8.29 MP) costs 125.10 and 87.54.
+A 1080p output (2.07 MP) costs 29.83 credits per second in `creative` mode and 20.88 in `precise`. A 4K output (8.29 MP) costs 119.34 and 83.54.
## Bria
@@ -163,6 +166,18 @@ A 1080p output (2.07 MP) costs 31.24 credits per second in `creative` mode and 2
| Bria Remove Video Background (Transparent) | 10.55 / sec |
| Bria Video Eraser | 13.58 / sec |
+## Breeze
+
+All Breeze voice nodes bill at **9.05 credits per 1K billable characters**, and CJK characters count as two. Text to Speech and Voice Design bill the text you enter. Voice Clone, Remix, and Localize use the same rate, but the character count is only known after the run: Remix and Localize bill the generated preview text, and Clone bills the vendor-reported billable units. The node badge shows an approximate envelope of **0.66–2.66**, **0.72–1.52**, and **0.76–2.29** credits for those three nodes.
+
+| Node | Credits |
+| :-------------------- | :-------------- |
+| Breeze Text to Speech | 9.05 / 1K chars |
+| Breeze Voice Design | 9.05 / 1K chars |
+| Breeze Voice Clone | 9.05 / 1K chars |
+| Breeze Voice Remix | 9.05 / 1K chars |
+| Breeze Voice Localize | 9.05 / 1K chars |
+
## ByteDance
### LLM: ByteDance Seed
@@ -183,15 +198,17 @@ Per successfully generated image.
| model | Credits / image |
| :-------------------------------------------------- | :------------------- |
+| seedream 5.5 (`seedream-5-5-261028`) | 6.03 (1K) / 10.56 (2K) |
| seedream 5.0 pro (`seedream-5-0-pro-260628`) | 9.50 (1K) / 18.99 (2K) |
| seedream 5.0 flash (`seedream-5-0-flash-260915`) | 5.43 |
| seedream 5.0 lite (`seedream-5-0-260128`) | 7.39 |
| seedream-4-5-251128 | 8.44 |
| seedream-4-0-250828 | 6.33 |
+| Seedream 5.5 Layer Separation | 6.03 (1K, 1.5K) / 10.56 (2K) |
| Seedream 5.0 Pro Layer Separation | 6.79 (1K, 1.5K) / 13.58 (2K) |
| Seedream 5.0 Flash Layer Separation | 5.43 |
-Seedream 5.0 Pro uses resolution-based billing: **1K** (≤2.61 MP, **9.50 credits**) and **2K** (>2.61 MP, **18.99 credits**). Seedream 5.0 Flash bills a flat **5.43 credits** per image at every resolution. The first image in each request is free. Each additional image beyond the first (including both output and input/reference images) adds **0.63 credits**.
+Seedream 5.0 Pro uses resolution-based billing: **1K** (≤2.61 MP, **9.50 credits**) and **2K** (>2.61 MP, **18.99 credits**). Seedream 5.0 Flash bills a flat **5.43 credits** per image at every resolution. The first image in each request is free. Each additional image beyond the first (including both output and input/reference images) adds **0.63 credits**. **Seedream 5.5** bills **6.03 credits** up to about 2.6 MP and **10.56 credits** above that, and adds **2.41 credits** for each connected reference image. **Seedream 5.5 Layer Separation** bills **6.03 credits** at 1K/1.5K and **10.56 credits** at 2K, per images in the run.
### Image: ByteDance Image (deprecated)
@@ -205,24 +222,45 @@ Seedream 5.0 Pro uses resolution-based billing: **1K** (≤2.61 MP, **9.50 cre
### Video: Seedance 2.5
-Token-based billing. Estimated token consumption = **(input video duration + output video duration) × output video width × output video height × output video frame rate ÷ 1024**. When input is text, image, or audio only, input video duration = 0. Actual usage from the API is authoritative.
+Token-based billing. The live ComfyUI price badge estimates tokens as **floor((output width × output height ÷ 1024) × (24 × billed duration + 1))**, then multiplies by the Credits / 1K tokens rate below (the `+ 1` is one extra frame). Billed duration is the output duration, plus the input video duration when a reference video is connected. When input is text, image, or audio only, input video duration = 0. Actual usage from the API is authoritative.
+
+| Model | Resolution | Video input | Credits / 1K tokens |
+| :----------------- | :--------- | :---------- | ------------------: |
+| Seedance 2.5 | 480p, 720p | no | 3.229 |
+| Seedance 2.5 | 480p, 720p | yes | 1.931 |
+| Seedance 2.5 | 1080p | no | 3.530 |
+| Seedance 2.5 | 1080p | yes | 2.112 |
+| Seedance 2.5 | 4k | no | 1.418 |
+| Seedance 2.5 | 4k | yes | 0.875 |
+| Seedance 2.5 Flash | 480p, 720p | no | 2.112 |
+| Seedance 2.5 Flash | 480p, 720p | yes | 1.297 |
+
+**Video input** means a reference video is connected. Text, image, and audio references count as no video input. The live badge does not apply a separate minimum-token floor beyond the formula above.
+
+**Seedance 2.5 Flash** generates 480p and 720p only, up to 30 seconds, and bills the same rate at both. A 5 second 720p Flash clip with no video input costs about 230 credits.
+
+A 4k frame carries about four times the tokens of a 1080p frame at the same duration, so the per-1K-token rate is lower while the total charge is higher.
-| Model | Resolution | Video input | Credits / 1K tokens |
-| :----------- | :--------- | :---------- | ------------------: |
-| Seedance 2.5 | 480p, 720p | no | 3.229 |
-| Seedance 2.5 | 480p, 720p | yes | 1.931 |
-| Seedance 2.5 | 1080p | no | 3.530 |
-| Seedance 2.5 | 1080p | yes | 2.112 |
+Reference videos must be at least 2 seconds long, and total input is limited to 30 seconds. Estimated charges for 16:9 output at 24 fps, using the token formula above:
-**Video input** means a reference video is connected. Text, image, and audio references count as no video input. Minimum token consumption limits apply when the input includes video: if the estimated consumption is below the limit, billing uses the minimum.
+| Resolution | 5s output, no video input | 5s output + 2s video input | 5s output + 30s video input |
+| :--------- | ------------------------: | -------------------------: | --------------------------: |
+| 480p | 156.38 | 130.64 | 650.12 |
+| 720p | 351.58 | 293.72 | 1461.63 |
+| 1080p | 865.00 | 722.82 | 3596.98 |
+| 4k | 1389.91 | 1197.81 | 5960.70 |
-Reference videos must be at least 2 seconds long, and total input is limited to 30 seconds. Estimated charges for 16:9 output at 24 fps:
+A connected reference video uses the lower per-1K-token rate, so 5 seconds plus a 2-second reference can cost less than 5 seconds with no video input. The live badge cannot see the input video length, so for a 5-second run with video attached it ranges from the output duration only (at the with-video rate) up to output plus 30 seconds.
+
+A run with the **Seedance 2.5 Draft** model renders a 480p preview at the 480p token rates in the first table above and returns a `draft_task_id`. **ByteDance Seedance 2.5 Draft to Final Video** renders the 1080p or 4k final of that draft; the final reuses the draft's prompt, references, duration and audio setting. The live node badge quotes a duration-independent credit range from the same formula (a 4-second 16:9 final with no reference video, up to a 30-second final plus 30 seconds of reference video). The high end is 21:9 at 1080p and 4:3 or 3:4 at 4k:
+
+| Node | Resolution | Credits |
+| :------------------------------------------ | :--------- | ----------------: |
+| ByteDance Seedance 2.5 Draft to Final Video | 1080p | 693.43 – 6202.64 |
+| ByteDance Seedance 2.5 Draft to Final Video | 4k | 1114.23 – 10214.09 |
+
+A 4 second 1080p final at 16:9 with no reference video costs 693.43 credits, and a 30 second 1080p final at 16:9 with 30 seconds of reference video costs 6163.19. Both totals include the badge's extra frame. A different aspect ratio costs a little more — the 1080p high end of 6202.64 is the 21:9 case, and the 4k high end of 10214.09 is the 4:3 or 3:4 case. Actual API usage determines the final charge.
-| Resolution | 5s output, no video input | 5s output + 2–30s video input |
-| :--------- | ------------------------: | ----------------------------: |
-| 480p | 155.1 | 166.9 – 649.3 |
-| 720p | 348.8 | 375.4 – 1459.8 |
-| 1080p | 857.8 | 923.9 – 3592.7 |
### Video: Seedance 2.0
480p and 720p share the same rate. Fast and Mini do not support 1080p or 4K.
@@ -262,7 +300,7 @@ Billed by tokens (same estimation formula as Seedance 2.0). Unlike Seedance 2.0,
| seedance-1-0-pro-fast-251015 | — | 211 |
- **Deprecated**: `seedance-1-5-pro-251215` retires on November 11, 2026, when BytePlus shuts the model down. ComfyUI removes it from the ByteDance 1.x nodes on the same date, and its pricing rows are removed from this page. Use `seedance-1-0-pro-250528`, `seedance-1-0-pro-fast-251015`, or a Seedance 2.x model instead.
+ **Retired**: BytePlus shut down `seedance-1-5-pro-251215` on November 11, 2026, so its pricing rows were removed from this page. Use `seedance-1-0-pro-250528`, `seedance-1-0-pro-fast-251015`, or a Seedance 2.x model instead.
**Already retired**: `seedance-1-0-lite-t2v-250428` and `seedance-1-0-lite-i2v-250428`, and the **ByteDance Reference Images to Video** node, have been retired and removed from ComfyUI. Their pricing below is retained for reference only.
@@ -354,7 +392,7 @@ At up to 30 fps, a 1080p output costs about 62.56 credits per second and the 9 M
## FishAudio
-**Limited-time offer**: from October 13 to November 12, 2026, Drama 3 is billed at about **3.02** credits per 1K bytes. The `drama-3` row in the table lists the standard price, which applies again from November 13, 2026.
+A launch rate of about **3.02** credits per 1K bytes for Drama 3 applied from October 13 to November 12, 2026. The `drama-3` row in the table lists the standard price, which applies from November 13, 2026.
| Node | Model | Credits |
@@ -397,6 +435,9 @@ Token-based billing; thought tokens use the output text rate.
| Model | Input credits / 1K | Output credits / 1K |
| :------------------------------------------------------ | -----------------: | :------------------ |
+| Gemini 4 Hydrogen (`gemini-4-hydrogen`) | 0.4526 | 2.263 |
+| Gemini 4 Argon (`gemini-4-argon`) | 1.20692 | 6.0346 |
+| Gemini 4 Carbon (`gemini-4-carbon`) | 1.20692 | 6.0346 |
| Gemini 3.8 Flash (`gemini-3.8-flash`) | 0.4526 | 2.263 |
| Gemini 3.7 Flash (`gemini-3.7-flash`) | 0.4526 | 2.263 |
| Gemini 3.5 Flash (`gemini-3.5-flash`) | 0.3165 | 1.899 |
@@ -428,6 +469,10 @@ Billed per second of generated video. Rates are approximate.
| | 720p | 30.57 |
| | 1080p | 45.87 |
| | 4k | 91.76 |
+| Gemini Omni Flash 1.2 (`gemini-omni-1.2-flash`) | 360p | 6.98 |
+| | 720p | 20.97 |
+| | 1080p | 31.46 |
+| | 4k | 62.92 |
### Video: Google Veo
@@ -483,23 +528,33 @@ Prices are set per-second, per-run, or per-1K characters.
| HeyGen Video Translate | speed mode | 4.07 / sec |
| HeyGen Text to Speech | N/A | ~0.20 / sec |
| HeyGen Video 1.0 Reference to Video | 480p, no references | 3.02 / sec |
-| HeyGen Video 1.0 Reference to Video | 768p, no references | 4.53 / sec |
-| HeyGen Video 1.0 Reference to Video | 2K, no references | 13.58 / sec |
-| HeyGen Video 1.0 Reference to Video | 480p, with reference media | 6.03 / sec |
-| HeyGen Video 1.0 Reference to Video | 768p, with reference media | 9.05 / sec |
-| HeyGen Video 1.0 Reference to Video | 2K, with reference media | 27.16 / sec |
+| HeyGen Video 1.0 Reference to Video | 768p, no references | 5.43 / sec |
+| HeyGen Video 1.0 Reference to Video | 1080p, no references | 12.07 / sec |
+| HeyGen Video 1.0 Reference to Video | 2K, no references | 15.09 / sec |
+| HeyGen Video 1.0 Reference to Video | 480p, with reference media | 10.56 / sec |
+| HeyGen Video 1.0 Reference to Video | 768p, with reference media | 15.09 / sec |
+| HeyGen Video 1.0 Reference to Video | 1080p or 2K, with reference media | 27.16 / sec |
+| HeyGen Video 1.0 Reference to Video | extra image after the first 2 | 6.03 / image |
| HeyGen Video 1.0 Image to Video | 480p | 3.02 / sec |
-| HeyGen Video 1.0 Image to Video | 768p | 4.53 / sec |
-| HeyGen Video 1.0 Image to Video | 2K | 13.58 / sec |
+| HeyGen Video 1.0 Image to Video | 768p | 5.43 / sec |
+| HeyGen Video 1.0 Image to Video | 1080p | 12.07 / sec |
+| HeyGen Video 1.0 Image to Video | 2K | 15.09 / sec |
+| HeyGen Video 1.0 Image to Video | over 15 seconds, 480p | 13.58 / sec |
+| HeyGen Video 1.0 Image to Video | over 15 seconds, 768p | 22.63 / sec |
+| HeyGen Video 1.0 Image to Video | over 15 seconds, 1080p or 2K | 31.68 / sec |
| HeyGen Voice 1.0 Instant Clone | N/A | Free |
| HeyGen Voice 1.0 Text to Speech | `heygen-voice-1` | 9.05 / 1K chars |
| HeyGen Voice 1.0 Text to Speech | `heygen-voice-1-turbo` | 2.26 / 1K chars |
-**Text to speech pricing**: a launch rate of about **4.53** credits per 1K characters for `heygen-voice-1` and **1.21** for `heygen-voice-1-turbo` applied through October 31, 2026. The rows above list the standard rates, which apply from November 1, 2026.
+**Text to speech pricing**: a launch rate of about **4.53** credits per 1K characters for `heygen-voice-1` and **1.21** for `heygen-voice-1-turbo` applied through November 1, 2026 (00:00 UTC). The rows above list the standard rates, which apply from that time.
Avatar video bills per second at the rate of the selected avatar look's type (Digital Twin, Photo Avatar, or Studio Avatar). **Auto** picks the engine per look, so its rate spans the full range above; a custom avatar ID is billed by its own type.
-HeyGen Video 1.0 clips run from 5 to 15 seconds and bill at the rate of the selected resolution. Reference to Video uses the no-reference rate until a reference image or video is connected, and each reference video adds 5 seconds to the billed duration. **2K** output requires the `16:9` or `9:16` aspect ratio; `auto` works at 2K only when no reference media is connected.
+**Limited-time offer**: a 50% promotional rate applies through December 1, 2026 (00:00 UTC). The Video 1.0 rows above show these current promotional rates, except the long-clip rates, which are already at the standard price. From December 1, 2026 (00:00 UTC), the standard rates apply: no-reference and Image to Video (≤15 seconds) **6.03 / 10.86 / 24.14 / 30.17** credits per second; with reference media, **21.12 / 30.17 / 54.31** credits per second; each extra image after the first two, **12.07 credits**. Long-clip rates stay as listed.
+
+HeyGen Video 1.0 clips run from 5 to 30 seconds and bill per second. Clips longer than 15 seconds use the long-clip rows for both nodes when no reference media is connected. Durations over 15 seconds are not available when references are connected. Connecting any image or video moves **Reference to Video** to the with-reference rate, even if the first two images have no per-image fee. The first two images are free; each image beyond the first two adds **6.03 credits** (that add-on becomes **12.07 credits** on December 1, 2026, 00:00 UTC). Up to 15 seconds of reference footage is billed at the with-reference per-second rate.
+
+`auto_duration` replaces `duration` with the length of the quoted dialogue: 8 to 30 seconds, or 8 to 15 seconds plus reference footage. **1080p** and **2K** output require the `16:9` or `9:16` aspect ratio; `auto` works at those resolutions only without reference media.
## Higgsfield
@@ -610,15 +665,15 @@ Total credits = **(credits / sec) × video duration**. Rates depend on input FPS
## Ideogram
-| Node | turbo | Character reference | rendering_speed | Credits / image |
-| :------------- | :---- | :------------------ | :-------------- | :-------------- |
-| Ideogram V1 | false | — | — | 18.1 |
-| Ideogram V1 | true | — | — | 6.03 |
-| Ideogram V2 | false | — | — | 24.14 |
-| Ideogram V2 | true | — | — | 15.09 |
-| Ideogram V4 | — | — | TURBO | 9.05 |
-| Ideogram V4 | — | — | DEFAULT | 18.1 |
-| Ideogram V4 | — | — | QUALITY | 30.17 |
+
+ **Deprecated**: ComfyUI removed the **Ideogram V1** and **Ideogram V2** nodes, so this page no longer lists their pricing. Use **Ideogram V4** instead.
+
+
+| Node | rendering_speed | Credits / image |
+| :---------- | :-------------- | :-------------: |
+| Ideogram V4 | TURBO | 9.05 |
+| Ideogram V4 | DEFAULT | 18.1 |
+| Ideogram V4 | QUALITY | 30.17 |
| Node | quality | resolution | Credits / image |
| :---------------- | :-------- | :--------- | :-------------- |
@@ -677,13 +732,41 @@ Total credits = **credits / image** at the selected `resolution` and `thinking_m
| Krea 3 | 2K | high | 24.14 | 6.03 - 21.12 |
| Krea 3 | 2K | ultra | 78.45 | 6.03 - 21.12 |
-At 2K, each reference image adds **6.03 to 21.12 credits**, depending on the reference image's pixel size; at 1K reference images are downscaled to about 1 MP, so the surcharge is a flat **6.03 credits**. `ultra` always outputs 2K regardless of `resolution` and requires at least one reference image.
+At 2K, each reference image adds **6.03–21.12 credits** (`6.03` at the low end, `21.12` at the high end; the node scales between these two bounds by the reference image's pixel size). At 1K, reference images are downscaled to about 1 MP, so the surcharge is a flat **6.03 credits**. `ultra` always outputs 2K regardless of `resolution` and requires at least one reference image.
## Kling
### Video: Credits / sec
-Total credits = **(credits / sec) × `duration`**, except **Kling 3.0 Omni Edit Video** and **Kling Motion Control** (billed per second of output; duration follows input video on Edit).
+Total credits = **(credits / sec) × `duration`**, except **Kling 3.0 Omni Edit Video**, **Kling 4.0 Edit Video** and **Kling Motion Control** (billed per second of output; duration follows input video on Edit).
+
+#### kling-4.0
+
+`kling-4.0` generates 3 to 30 seconds at 720p or 1080p. `kling-4.0-flash` generates 3 to 20 seconds at 720p.
+
+| Node | model | resolution | Credits / sec | Credits / sec, reference videos |
+| :--------------------------- | :-------------- | :--------- | :------------ | :------------------------------ |
+| Kling 4.0 Text to Video | kling-4.0 | 720p | 47.37 | — |
+| Kling 4.0 Text to Video | kling-4.0 | 1080p | 70.91 | — |
+| Kling 4.0 Image to Video | kling-4.0 | 720p | 47.37 | — |
+| Kling 4.0 Image to Video | kling-4.0 | 1080p | 70.91 | — |
+| Kling 4.0 Reference to Video | kling-4.0 | 720p | 47.37 | 28.36 |
+| Kling 4.0 Reference to Video | kling-4.0 | 1080p | 70.91 | 42.54 |
+| Kling 4.0 Text to Video | kling-4.0-flash | 720p | 28.36 | — |
+| Kling 4.0 Image to Video | kling-4.0-flash | 720p | 28.36 | — |
+| Kling 4.0 Reference to Video | kling-4.0-flash | 720p | 28.36 | 19.01 |
+
+On **Kling 4.0 Reference to Video**, connected reference videos switch the whole run to the lower rate: every second of output and every second of reference bills at that rate. Each clip is 3 to 30 seconds (3 to 20 on `kling-4.0-flash`) and all clips together stay within the same limit. A 5 second 720p run with 30 seconds of reference video costs 992.7 credits.
+
+**Kling 4.0 Edit Video** always bills at the lower (reference-video) rate: **credits = rate × (input video seconds + reference video seconds)**. The output is the same length as the input, so with no reference videos this is **2 × rate** per second of input. With reference videos, the actual charge follows that same formula; the live badge shows **1× to 2× rate** per second of input because it cannot see the video lengths. Reference images add no per-image fee.
+
+| model | resolution | Rate (credits / sec) | No reference videos | With reference videos |
+| :-------------- | :--------- | -------------------: | :------------------ | :-------------------- |
+| kling-4.0 | 720p | 28.36 | 56.73 / sec of input (2 × rate) | 28.36 × (input + ref seconds) |
+| kling-4.0 | 1080p | 42.54 | 85.09 / sec of input (2 × rate) | 42.54 × (input + ref seconds) |
+| kling-4.0-flash | 720p | 19.01 | 38.02 / sec of input (2 × rate) | 19.01 × (input + ref seconds) |
+
+A 10 second 1080p edit without reference videos costs 850.9 credits.
#### kling-video-o1
@@ -739,10 +822,6 @@ Always generates audio; the audio toggle is ignored.
| motion control | — | std | 14.77 |
| motion control | — | pro | 23.63 |
-
- **End of life (EOL)**: Kling V1.5, V1.6, V2.1, V2.1 Master, and Kolors Virtual Try-On are scheduled for retirement on **September 15, 2026**. The pricing below is retained for reference until removal.
-
-
#### kling-v2-5-turbo
**Kling Text to Video**, **Kling Image (First Frame) to Video**, and **Kling Start-End Frame to Video** use `kling-v2-5-turbo` in **pro** mode only, billed per run at a fixed 5s or 10s duration.
@@ -790,6 +869,10 @@ Per run.
| :------------------- | :------------ |
| Kling Virtual Try-On | 33.79 |
+
+ **Retired**: Kling retired V1.5, V1.6, V2.1, V2.1 Master, and the earlier Kolors Virtual Try-On service on **September 15, 2026**. The Kling nodes no longer offer those models, so their pricing rows were removed from this page. The **Kling Virtual Try-On** row above is a different, current node billed per run.
+
+
## Lightricks
### Video: Credits / sec
@@ -819,15 +902,35 @@ Total credits = **(credits / sec) × `duration`**.
Resolution maps to `1280x720` / `720x1280` (720p), `1920x1080` / `1080x1920` (1080p), `2560x1440` / `1440x2560` (1440p), and `3840x2160` / `2160x3840` (4K). Durations over 10s on **LTX-2.5 (Fast)** require a 720p or 1080p resolution and 24/25 FPS; on **LTX-2 (Fast)** they require 1080p and 25 FPS.
+### Post-processing: Credits / sec
+
+These nodes process the input video and bill per second of it.
+
+| Node | resolution | Credits / sec |
+| :------------------- | ---------- | --------------- |
+| LTX Video AlphaGen | 1080p (short side ≤1080; below 1080 uses this tier) | 30.17 |
+| LTX Video AlphaGen | 1440p (short side ≤1440) | 75.43 |
+| LTX Video AlphaGen | 2160p (short side ≤2160) | 301.73 |
+| LTX Video SDR to HDR | 1080p (≤ 1920×1080) | 120.69 |
+| LTX Video SDR to HDR | 1440p (≤ 2560×1440) | 241.39 |
+| LTX Video SDR to HDR | 4K (≤ 3840×2160) | 482.77 |
+| LTX Video Reframe | 720p | 30.17 |
+| LTX Video Reframe | 1080p | 60.35 |
+| LTX Video Upscale | 1080p | 24.14 |
+| LTX Video Upscale | 4k | 90.52 |
+| LTX Video Upscale | 8k | 362.08 |
+
+**AlphaGen** is billed per second of input video, by the input's short side (the smaller of width and height). Inputs with a short side below 1080 pixels use the 1080p rate. **SDR to HDR** is billed per second of input video; the input snaps to the smallest resolution tier that can contain its pixel count. The live ComfyUI badge cannot see the input resolution, so it shows only the envelope (**30.17–301.73** credits/sec for AlphaGen, **120.69–482.77** for SDR to HDR). The actual charge uses the three-tier rates in the table. **AlphaGen** takes up to 10 seconds of video, **Reframe** up to 60 seconds, and **SDR to HDR** and **Upscale** cap the frame count by resolution (273 frames at 1080p for SDR to HDR, 2,401 for Upscale).
+
## Luma
### Video
- **Deprecated**: The legacy **Luma Text to Video** and **Luma Image to Video** nodes are deprecated, and Luma removed the endpoints they use on October 24, 2026, so these two nodes can no longer generate videos. Use the **Luma Ray 3.2** nodes instead. The ray-2, ray-flash-2, and ray-1-6 rows below are retained for reference only; **ray-1-6** stopped being served in September 2026.
+ **Retired**: the legacy Luma Text to Video, Image to Video and image generation nodes are removed, along with the ray-2, ray-flash-2, ray-1-6, photon-1 and photon-flash-1 models. The tables below list only the nodes ComfyUI still ships.
-Flat **/run** prices match the node badge for each `duration` choice. **Edit** and **reframe** bill **(credits / sec) × video length**. **ray-1-6** is always **105.50 / run**.
+Flat **/run** prices match the node badge for each `duration` choice. **Edit** and **reframe** bill **(credits / sec) × video length**.
| Model | task | resolution | Credits |
| :---------- | :--------- | :--------- | :------------------------------------------------ |
@@ -843,31 +946,18 @@ Flat **/run** prices match the node badge for each `duration` choice. **Edit** a
| ray-3.2 | reframe | 540p | 12.66 / sec |
| ray-3.2 | reframe | 720p | 25.32 / sec |
| ray-3.2 | reframe | 1080p | 75.96 / sec |
-| ray-2 | generation | 540p | 120.27 / run (5s), 217.33 / run (9s) |
-| ray-2 | generation | 720p | 215.22 / run (5s), 386.13 / run (9s) |
-| ray-2 | generation | 1080p | 478.97 / run (5s), 865.10 / run (9s) |
-| ray-2 | generation | 4K | 1922.21 / run (5s), 3460.40 / run (9s) |
-| ray-flash-2 | generation | 540p | 42.20 / run (5s), 75.96 / run (9s) |
-| ray-flash-2 | generation | 720p | 71.74 / run (5s), 128.71 / run (9s) |
-| ray-flash-2 | generation | 1080p | 166.69 / run (5s), 299.62 / run (9s) |
-| ray-flash-2 | generation | 4K | 660.43 / run (5s), 1192.15 / run (9s) |
-| ray-1-6 | generation | — | 105.50 / run |
-
-**Luma Ray 3.2** Image to Video and Extend Video always bill the **5s** price. Legacy Image to Video uses **5s** or **9s** from **ray-2** / **ray-flash-2** rows.
+
+**Luma Ray 3.2** Image to Video and Extend Video always bill the **5s** price.
### Image
| Model | task | Credits |
| :------------- | :--------- | :----------------------------------- |
-| photon-flash-1 | — | 0.57 / run |
-| photon-1 | — | 2.19 / run |
| uni-1 | generation | 8.52 / run + 0.63 / ref image (≤ 9) |
| uni-1 | edit | 9.16 / run + 0.63 / ref image (≤ 8) |
| uni-1-max | generation | 21.10 / run + 0.63 / ref image (≤ 9) |
| uni-1-max | edit | 21.73 / run + 0.63 / ref image (≤ 8) |
-Reference-image surcharge uses USD rounded to 4 decimals before converting to credits.
-
## Meshy
| Node | Options | Credits |
@@ -926,7 +1016,7 @@ Total credits = **(credits / sec) × seconds billed**. **Mirelo Video to SFX** b
| MiniMax H3 First-Last-Frame to Video | MiniMax H3 | 2K | 39.22 / sec |
| MiniMax H3 Reference to Video | MiniMax H3 | 768P | 27.16 / sec, plus 12.07 / image over 5 |
| MiniMax H3 Reference to Video | MiniMax H3 | 2K | 39.22 / sec, plus 12.07 / image over 5 |
-| MiniMax H3 Context IR | MiniMax H3 | — | 4.22 / run (text only), 10.55 / run (with first/last frame), 12.66–23.21 / run (with reference media) |
+| MiniMax H3 Context IR | MiniMax H3 | — | 271.56 / 1M input tokens + 1,086.23 / 1M output tokens |
| MiniMax H3 Regenerate to 2K | MiniMax H3 | — | 15.09 / sec |
| MiniMax H3 Text to Video | MiniMax H3 Max | 480P | 15.09 / sec |
| MiniMax H3 Text to Video | MiniMax H3 Max | 768P | 24.14 / sec |
@@ -959,7 +1049,7 @@ Total credits = **(credits / sec) × seconds billed**. **Mirelo Video to SFX** b
**MiniMax Hailuo 02 Video** only supports **6s** at 1080P.
-**MiniMax H3** bills per second (4–15s duration). **MiniMax H3 Reference to Video** adds 12.07 credits per reference image beyond 5, and bills reference videos at 27.16 credits/sec (768P) or 39.22 credits/sec (2K) (2s per video, up to 15s total, approximate). **MiniMax H3 Context IR** prices depend on attached media: text only, first/last frame, or reference media (approximate).
+**MiniMax H3** bills per second (4–15s duration). **MiniMax H3 Reference to Video** adds 12.07 credits per reference image beyond 5, and bills reference videos at 27.16 credits/sec (768P) or 39.22 credits/sec (2K) (2s per video, up to 15s total, approximate). **MiniMax H3 Context IR** is billed at **271.56 credits per 1M input tokens** and **1,086.23 credits per 1M output tokens**. The node badge cannot see the token count at graph time, so it shows an approximate envelope of about **4.22** (text only), **10.55** (with first/last frame), or **12.66–23.21** (with reference media).
**MiniMax H3 Max** bills per second at the rates above. **Text to Video**, **First-Last-Frame to Video**, and **Reference to Video** use **5–15s** output duration. **Lip Sync** accepts audio from 5 to ~899 seconds; when output exceeds 15 seconds, the **1.2×** multiplier below applies.
@@ -985,6 +1075,19 @@ Worked examples, each assuming no other reference input consumes the free allowa
**MiniMax H3 Max Recast**, **Extend Video**, **Reframe Video**, **Insert Video**, **Lip Sync** and **Camera Control** bill per second of video. Extend Video also runs on the Turbo model at half the Max rate. **Reframe Video** is billed at the table rate per second. Longer sources use overlapping windows, so billed seconds can exceed the final output duration. **Lip Sync** is billed at the table rate per second. When the output is **over 15 seconds**, a **1.2×** multiplier applies (for example, 768P → 28.97 / sec). Insert Video adds reference tokens beyond the free 4,096-token allowance at 6.03 credits per 1K tokens, the same allowance as Reference to Video: 240 base tokens, 2,048 per image and 225,000 per reference video at 768P output (87,000 at 480P).
+### Image: MiniMax H3 Image
+
+Per generated image. The charge combines the output size with the prompt and reference images, which bill as tokens.
+
+| Node | model | size | Credits / image |
+| :--------------- | -------------------------- | ---- | --------------- |
+| MiniMax H3 Image | MiniMax-H3-Image | 2K | 5.32 – 8.17 |
+| MiniMax H3 Image | MiniMax-H3-Image | 4K | 19.41 – 22.26 |
+| MiniMax H3 Image | MiniMax-H3-Image-Highspeed | 2K | 10.02 – 12.86 |
+| MiniMax H3 Image | MiniMax-H3-Image-Highspeed | 4K | 38.20 – 41.04 |
+
+The table is the text-only, `enhance_prompt` on slice of this formula. Output credits = `max(pixels, 4,194,304) / 1,024 × 0.00114657` (use `0.00229315` on Highspeed): **4.70 credits** at 2K (4 MP) and about **18.79 credits** at 4K, or **9.39** / **37.57** on Highspeed. Prompt tokens bill at **0.000694 credits** each: **0–300** with no references and enhancement off, **900–5,000** with no references and enhancement on, or **2,000 + 300 × (n − 1)** to **6,500 + 800 × (n − 1)** with `n` reference images. Each reference image also adds **2,050–8,194** image tokens at **0.000453 credits** each (**0.000905** on Highspeed), or about **0.93 to 3.71 credits** (**1.86 to 7.42** on Highspeed). A 2K image with enhancement off costs about **4.70 credits**. A custom size bills its own area with a floor of 4 MP.
+
## OpenAI
DALL·E and Sora use fixed per-image or per-second rates. **OpenAI GPT Image 2** and **OpenAI ChatGPT** bill from API token usage.
@@ -1035,14 +1138,55 @@ The **GPT Image 2.5** models (`gpt-image-2.5-flare`, `gpt-image-2.5-sunburst`) s
Output token counts scale with size and quality (for example, 1024x1024 at `low` is about 196 output tokens, which is about 1.32 credits on `gpt-image-1.5`). Each 1024x1024 reference image adds about 2.47 credits of image input, and prompt text is billed at the text-input rate. Multiplying images (`n`) scales the total.
+The **GPT Image 2.6** models (`gpt-image-2.6-flare`, `gpt-image-2.6-sunburst`) share one rate table. Credits are per image at 1024x1024:
+
+| quality | Credits / image |
+| :------ | :-------------- |
+| low | 7.45 |
+| medium | 13.12 |
+| high | 15.89 |
+| xhigh | 28.25 |
+| max | 63.57 |
+
+Larger presets cost more: 3840x2160 is 14.12 credits at `low` and 120.8 at `max`. Total credits = **(output price + references × 2.47 to 3.71) × `n`**. Each reference image adds **2.47–3.71 credits**. The live badge cannot see the reference size that sets the exact per-reference charge, so it shows that envelope; `n` multiplies the total.
+
+### Image: GPT Image 2.6 GIF
+
+**OpenAI GPT Image 2.6 GIF** bills per run, not per token. Every frame is a full image, so frame size, quality, and `frame_count` set the price. `frame_count` bills in steps of 2 to 8, 9, 10 to 12, and 13 to 16 frames, and the largest sizes only fit the low steps. Credits for a 512x512 run of 9 frames:
+
+| quality | Credits / run |
+| :------ | :------------ |
+| low | 10.61 |
+| medium | 18.76 |
+| high | 22.72 |
+| xhigh | 40.41 |
+| max | 90.92 |
+
+A 384x384 run of 2 to 8 frames costs 3.88 credits at `low`, and a 720x720 run of 13 to 16 frames costs 214.69 at `max`. `frame_duration_ms` does not change the price.
+
+### Image: GPT Image 2.6 SVG
+
+**OpenAI GPT Image 2.6 SVG** bills the still-image price for the selected size and quality, plus **0.23–1.77 credits** for the vector pass. The live badge cannot see the input that sets the exact vector-pass charge, so it shows that envelope on top of the still-image price. The table ranges are the still-image price plus that envelope. The SVG badge does not add a separate reference-image charge. Sizes are 1024x1024, 1024x1536, and 1536x1024. Credits at 1024x1024:
+
+| quality | Credits / run |
+| :------ | :------------ |
+| low | 7.68 – 9.22 |
+| medium | 13.36 – 14.90 |
+| high | 16.12 – 17.66 |
+| xhigh | 28.48 – 30.03 |
+| max | 63.81 – 65.35 |
+
+A 1024x1536 or 1536x1024 run is cheaper: 6.25 to 7.79 credits at `low`, and 49.90 to 51.44 at `max`.
+
### Chat
Token-based billing. Badge rates are **per 1K tokens**; table is **per 1M tokens**.
| Model | Input credits / 1M | Output credits / 1M |
| :----------- | -----------------: | :------------------ |
+| gpt-6.1-sol, gpt-6-sol | 603.46 | 3017.3 |
+| gpt-6.1-terra | 301.73 | 1508.65 |
| gpt-6-astra | 3017.3 | 15086.5 |
-| gpt-6-sol | 603.46 | 3017.3 |
| gpt-6-luna | 30.17 | 150.87 |
| gpt-5.6-sol, gpt-5.6-sol-pro | 1055 | 6330 |
| gpt-5.6-terra, gpt-5.6-terra-pro | 527.5 | 3165 |
@@ -1102,10 +1246,12 @@ Badge rates are **per 1K tokens**; table is **per 1M tokens**. Final charges fol
| x-ai/grok-4.5 | 603.46 | 1810.38 |
| x-ai/grok-4.20 | 377.16 | 754.33 |
| x-ai/grok-4.3 | 377.16 | 754.33 |
+| meta/muse-wonder | 1206.92 | 6034.60 |
+| meta/muse-spark-1.3 | 377.16 | 1282.35 |
+| meta/muse-ember | 30.17 | 150.87 |
| deepseek/deepseek-v4-pro | 131.25 | 262.51 |
| deepseek/deepseek-v4-flash | 33.79 | 67.59 |
| deepseek/deepseek-v3.2 | 76.04 | 114.05 |
-| qwen/qwen3.6-max-preview | 313.80 | 1882.80 |
| qwen/qwen3.6-plus | 98.06 | 588.37 |
| qwen/qwen3.6-flash | 56.57 | 339.45 |
| mistralai/mistral-large-2512 | 150.87 | 452.60 |
@@ -1147,6 +1293,39 @@ All three nodes bill per run, and every PBR map you enable adds to the price. `M
Prices are approximate. A 1 MP material with the default `basecolor` and `normal` maps costs 15.09 credits, and a 4 MP output with all five maps at 4× upscale costs 184.06 credits for **Patina Material** or 211.21 credits for **Patina Material Extract**.
+## Pika
+
+Pika H3 Flex bills per second of video. Total credits = **(credits / sec) × seconds billed**. `duration` is 5 to 15 seconds. At `480p` and `768p` the billed length equals `duration`. From `1080p` up the node fits the clip to its frame grid, so a 5 second run bills 5.17 seconds, a 10 second run bills 10.13, and a 15 second run bills 15.08.
+
+| Node | resolution | Credits / sec |
+| :------------------------------ | :--------- | :------------ |
+| Pika H3 Flex Video | 480p | 3.02 |
+| Pika H3 Flex Video | 768p | 6.03 |
+| Pika H3 Flex Video | 1080p | 10.56 |
+| Pika H3 Flex Video | 2k | 12.07 |
+| Pika H3 Flex Video | 4k | 24.14 |
+| Pika H3 Flex Reference to Video | 480p | 9.05 |
+| Pika H3 Flex Reference to Video | 768p | 18.10 |
+| Pika H3 Flex Reference to Video | 1080p | 22.63 |
+| Pika H3 Flex Reference to Video | 2k | 24.14 |
+| Pika H3 Flex Reference to Video | 4k | 30.17 |
+
+A 5 second 480p run of **Pika H3 Flex Video** costs 15.09 credits, and a 5 second 1080p run costs 54.56.
+
+**Pika H3 Flex Reference to Video** bills connected reference videos at the same rate. Each clip runs 2 to 15 seconds and reference video time is capped at 15 seconds, so the charge depends on how much reference video you connect. At 480p, a 5 second run with one 2 second reference costs 63.36 credits. At 4k, a 5 second run with 15 seconds of reference video costs 608.49. All Pika examples use the unrounded per-second rates.
+
+### Audio: Pika
+
+The Pika audio nodes also bill per second. **Pika Soundtrack** charges the input video's length (up to 10 minutes), **Pika SFX** charges `duration` (1 to 20 seconds), and **Pika Music** charges the track: `duration` in `text to music` (10 to 360 seconds), or the source audio length clamped to 5 to 360 seconds in `cover`.
+
+| Node | Credits / sec |
+| :-------------- | :------------ |
+| Pika Soundtrack | 1.509 |
+| Pika SFX | 0.060 |
+| Pika Music | 0.075 |
+
+A 10 minute Soundtrack run costs 905.2 credits, a 20 second SFX run 1.21, and a 3 minute Music track 13.58.
+
## Pixverse
Model **v3.5** (fixed in the API; not selectable).
@@ -1173,19 +1352,12 @@ PixVerse V6 nodes generate videos from 1 to 15 seconds with optional native audi
| PixVerse V6 | 540p | 7.18 | 9.23 |
| PixVerse V6 | 720p | 9.23 | 12.31 |
| PixVerse V6 | 1080p | 18.47 | 23.60 |
-| PixVerse V6 Max | 360p | 6.71 | 9.39 |
-| PixVerse V6 Max | 540p | 9.39 | 12.07 |
-| PixVerse V6 Max | 720p | 12.07 | 16.09 |
-| PixVerse V6 Max | 1080p | 24.14 | 30.84 |
-
-**PixVerse V6 Max** runs at the rates above as an introductory price. From **November 3, 2026** the per-second rates are:
+| PixVerse V6 Max | 360p | 8.05 | 12.07 |
+| PixVerse V6 Max | 540p | 12.07 | 14.75 |
+| PixVerse V6 Max | 720p | 14.75 | 20.12 |
+| PixVerse V6 Max | 1080p | 29.50 | 37.55 |
-| quality | Audio off | Audio on |
-| :------ | :-------- | :------- |
-| 360p | 8.05 | 12.07 |
-| 540p | 12.07 | 14.75 |
-| 720p | 14.75 | 20.12 |
-| 1080p | 29.50 | 37.55 |
+**PixVerse V6 Max** ran at lower introductory rates through **November 2, 2026**. The rows above list the standard rates, which apply from **November 3, 2026**.
**PixVerse V6 Fusion** bills 2x the per-second rate for each second of reference video, instead of the output duration. Fusion runs on `PixVerse V6` rates only.
@@ -1193,21 +1365,10 @@ PixVerse V6 nodes generate videos from 1 to 15 seconds with optional native audi
### Image
-The retired **arrow-1.1** models charged a fixed amount per run:
-
- **Deprecated**: Quiver retired the Arrow 1-series models on **October 16, 2026**, and the Text to SVG and Image to SVG nodes no longer list `arrow-1.1`, `arrow-1.1-max`, or `arrow-preview`. Use the Arrow 2 models instead. The pricing below is retained for reference only.
+ **Retired**: Quiver retired the Arrow 1-series models (`arrow-1.1`, `arrow-1.1-max`, `arrow-preview`) on **October 16, 2026**, and the Text to SVG and Image to SVG nodes no longer list them. Use the Arrow 2 models below.
-| model | Node | Credits / run |
-| :------------ | :------------------ | :---------------- |
-| arrow-1.1 | Quiver Text to SVG | 60.35 |
-| arrow-1.1 | Quiver Image to SVG | 45.26 |
-| arrow-1.1-max | Quiver Text to SVG | 75.43 |
-| arrow-1.1-max | Quiver Image to SVG | 60.35 |
-| arrow-preview | Quiver Text to SVG | 90.52 |
-| arrow-preview | Quiver Image to SVG | 90.52 |
-
**arrow-2** and **arrow-2-telos** bill by the tokens a request consumes:
| model | Input credits / 1M tokens | Output credits / 1M tokens |
@@ -1328,12 +1489,15 @@ Gen4 Turbo: **5s** or **10s**. Total credits = **(credits / sec) × `duration`**
## Sonilo
-Text to Music: total credits = **(credits / sec) × `duration`** (1–360 seconds).
+Video to Music bills per second of input video, up to 6 minutes. Text to Music and Music 1.5 bill per second of audio: total credits = **(credits / sec) × `duration`**.
| Node | Credits |
| :-------------------- | :----------- |
-| Sonilo Video to Music | 1.9 / sec |
-| Sonilo Text to Music | 0.5275 / sec |
+| Sonilo Video to Music | 2.72 / sec |
+| Sonilo Text to Music | 0.6789 / sec |
+| Sonilo Music 1.5 | 0.6789 / sec |
+
+Text to Music renders a duration from 1 to 360 seconds. Music 1.5 fits the song to the lyrics with `auto`, or renders the `custom` length you set, from 10 to 360 seconds.
## sync.so
@@ -1404,7 +1568,7 @@ Multiview applies only to **Hunyuan3D: Image(s) to Model** when `image_left`, `i
| :-------- | :----------- |
| Reimagine | 67.52 / 8MP |
| Bloom 2 | 96.55 / 8MP |
-| Wonder 3.5 | 24.14 / 8MP |
+| Wonder 3.5, Wonder 4 | 24.14 / 8MP |
**Topaz Image Enhance (Legacy)** is deprecated. It has no fixed pre-run estimate; charge comes from the API estimate at poll time (API credits × 0.08 USD). Typical charges by input resolution:
@@ -1564,6 +1728,19 @@ Pruna P-Video-2 bills per second of generated video. Total credits = **(credits
| | 1080p | off | 15.09 |
| | 1080p | on | 9.05 |
+### Video: P-Video Upscale
+
+Pruna P-Video Upscale bills per second of output video. Total credits = **(credits / sec) × output length**. The rate follows `target`, output above 30 fps costs twice the rate, and a source larger than the target keeps its size.
+
+A launch rate of 1.51, 3.02, 6.03 and 12.07 credits per second for `720p`, `1080p`, `2K` and `4K` applied through November 12, 2026 (17:00 UTC). The rows below list the standard rates, which apply from that date.
+
+| target | Target size | Credits / sec |
+| :----- | :----------- | :------------ |
+| 720p | about 0.9 MP | 3.02 |
+| 1080p | about 2.1 MP | 6.03 |
+| 2K | about 3.7 MP | 12.07 |
+| 4K | about 8.3 MP | 24.14 |
+
## VEED
| Node | Options | Credits |
@@ -1653,23 +1830,25 @@ Fixed **5s** at **1080p**.
| wan2.5-t2i-preview, wan2.5-i2i-preview | — | 6.33 / run |
| wan2.7-t2v, wan2.7-i2v | 720P | 21.1 / sec |
| | 1080P | 31.65 / sec |
-| wan2.7-i2v (continuation) | 720P | 21.1 / sec (range) |
-| | 1080P | 31.65 / sec (range) |
+| wan2.7-i2v (continuation) | 720P | 21.1 / sec (output + 2–5s input) |
+| | 1080P | 31.65 / sec (output + 2–5s input) |
| wan2.7-videoedit | 720P | 21.1 / sec (in + out) |
| | 1080P | 31.65 / sec (in + out) |
-| wan2.7-r2v | 720P | 21.1 / sec (range) |
-| | 1080P | 31.65 / sec (range) |
-| wan3.0-video | 480P | 15.09 / sec (range) |
-| | 720P | 30.17 / sec (range) |
-| | 1080P | 60.35 / sec (range) |
-| wan3.0-video-prime | 480P | 20.52 / sec (range) |
-| | 720P | 42.24 / sec (range) |
-| | 1080P | 84.48 / sec (range) |
+| wan2.7-r2v | 720P | 21.1 / sec (output to output + 5s input) |
+| | 1080P | 31.65 / sec (output to output + 5s input) |
+| wan3.0-video | 480P | 15.09 / sec |
+| | 720P | 30.17 / sec |
+| | 1080P | 60.35 / sec |
+| wan3.0-video-prime | 480P | 20.52 / sec |
+| | 720P | 42.24 / sec |
+| | 1080P | 84.48 / sec |
| wan2.5-t2v-preview, wan2.5-i2v-preview | 480p | 10.55 / sec |
| wan2.5-t2v-preview, wan2.5-i2v-preview, wan2.6-t2v, wan2.6-i2v | 720p | 21.1 / sec |
| wan2.5-t2v-preview, wan2.5-i2v-preview, wan2.6-t2v, wan2.6-i2v | 1080p | 31.65 / sec |
-| wan2.6-r2v | 720p | 21.1 / sec (range) |
-| | 1080p | 31.65 / sec (range) |
+| wan2.6-r2v | 720p | 21.1 / sec (output + 2–5s input) |
+| | 1080p | 31.65 / sec (output + 2–5s input) |
+
+**Wan 2.6 Reference to Video** and **Wan 2.7 continuation** bill the output duration plus **2–5 seconds** of input, at the same per-second rate. **Wan 2.7 Reference to Video** ranges from the output duration only up to the output duration plus **5 seconds** of input. **Wan 2.7 Edit** bills input plus output at the table rate. **Wan 3.0** bills at the table rate per second: if `duration` is `auto`, the badge shows that per-second rate; otherwise it ranges from the selected duration up to `min(duration + 15, 30)` seconds.
## WaveSpeed
@@ -1694,17 +1873,23 @@ Video endpoints charge for moderated content.
| :------------------------- | :--------- | :------- | :------------------------------------------- |
| grok-imagine-image-2.0 | 1K | generate | 12.66 (medium) / 8.44 (low) per output image |
| | 2K | generate | 16.88 (medium) / 12.66 (low) per output image |
-| grok-imagine-image-quality | 1K | generate | 10.55 / output image |
-| | 2K | generate | 14.77 / output image |
-| grok-imagine-image-pro | — | generate | 14.77 / output image |
+| grok-imagine-image-quality | 1K | generate | 8.44 / output image |
+| | 2K | generate | 12.66 / output image |
+| grok-imagine-image-pro | 1K | generate | 8.44 / output image |
+| | 2K | generate | 12.66 / output image |
| grok-imagine-image | — | generate | 4.22 / output image |
-| grok-imagine-image-2.0 | 1K | edit | 2.11 input + 12.66 (medium) / 8.44 (low) per output image |
-| | 2K | edit | 2.11 input + 16.88 (medium) / 12.66 (low) per output image |
-| grok-imagine-image-quality | 1K | edit | 2.11 input + 10.55 / output image |
-| | 2K | edit | 2.11 input + 14.77 / output image |
-| grok-imagine-image-pro | — | edit | 2.11 input + 14.77 / output image |
+| grok-imagine-image-2.0 | 1K | edit | 2.11–6.33 input + 12.66 (medium) / 8.44 (low) per output image |
+| | 2K | edit | 2.11–6.33 input + 16.88 (medium) / 12.66 (low) per output image |
+| grok-imagine-image-quality | 1K | edit | 2.11–6.33 input + 8.44 / output image |
+| | 2K | edit | 2.11–6.33 input + 12.66 / output image |
+| grok-imagine-image-pro | 1K | edit | 2.11 input + 8.44 / output image |
+| | 2K | edit | 2.11 input + 12.66 / output image |
| grok-imagine-image | — | edit | 0.422–1.266 input + 4.22 / output image |
+For non-pro edit models, the input charge is **2.11 credits** per input image for **1 to 3** images (**2.11–6.33**). `grok-imagine-image` uses **0.422 credits** per input image (**0.422–1.266**). The pro edit rows bill exactly one input image at **2.11 credits**.
+
+`grok-imagine-image-quality` and `grok-imagine-image-pro` are retired by xAI. The image nodes still list the two models and bill them at the `grok-imagine-image-2.0` low-quality rate.
+
### Video
| model | resolution | Credits |
@@ -1722,6 +1907,8 @@ Video endpoints charge for moderated content.
| | 720p | 21.12 / sec output + 0.603 / ref image |
| grok-imagine-video (extend)| — | 6.03–45.26 input + 15.09 × extension sec output / run |
+**Extend** bills **15.09 credits** per second of extension, plus an input charge of **6.03 credits** for a 2-second source video or **45.26 credits** for a 15-second source (the source must be 2–15 seconds). The live badge cannot see the source length, so it shows that **6.03–45.26** input band plus the per-second extension charge.
+
## Cloud GPU
| gpu_type | Credits |
diff --git a/zh/tutorials/partner-nodes/pricing.mdx b/zh/tutorials/partner-nodes/pricing.mdx
index de62fd3ba7..483a53644e 100644
--- a/zh/tutorials/partner-nodes/pricing.mdx
+++ b/zh/tutorials/partner-nodes/pricing.mdx
@@ -3,54 +3,56 @@ title: "Partner Node 定价"
description: "各提供商(包括 Anthropic、OpenAI、Kling 和 Luma)的 Partner Node 积分定价,按模型、Token 类型和生成类型细分。"
sidebarTitle: "定价"
mode: "wide"
-translationSourceHash: 2ef8ce31
+translationSourceHash: f68d51fd
translationFrom: tutorials/partner-nodes/pricing.mdx
translationBlockHashes:
"_intro": 1ec20604
- "Anthropic": 61f55fec
+ "Anthropic": 1757fac0
"Beeble": d1b19f61
- "BFL": 7ae11dc6
+ "BFL": 0a63f1b4
"Bria": 94080005
- "ByteDance": d2af4769
+ "Breeze": f38692c6
+ "ByteDance": fec72d04
"Clarity AI": 3d568b52
"ElevenLabs": cadb48cc
- "FishAudio": c957b7da
+ "FishAudio": 24ec3ea0
"Magnific": 4000e7e1
- "Google": b8be5b63
+ "Google": 958913c2
"HappyHorse": bceb584a
- "HeyGen": 7f08ab4e
+ "HeyGen": 248354a7
"Higgsfield": eed76dcb
"Hitpaw": 320c0537
- "Ideogram": da793624
+ "Ideogram": de0981df
"Krea": bcad7ba2
- "Kling": 32a38b1d
- "Lightricks": 3ba8d019
- "Luma": 9f809a47
+ "Kling": db91f830
+ "Lightricks": bbcb475e
+ "Luma": a2f2a8de
"Meshy": c057a289
"Meta": a253fe9a
"Mirelo": 2055c18a
- "Minimax": 07381eba
- "OpenAI": 17e24f34
- "OpenRouter": ef37d425
+ "Minimax": 3f9b84c5
+ "OpenAI": e51a94f8
+ "OpenRouter": 3de75ece
"Patina": d489c37e
- "Pixverse": bdb9b1c9
- "Quiver": 5538ce9c
+ "Pika": 66ba8620
+ "Pixverse": 86f21b3f
+ "Quiver": f84ffa99
"Qwen": 22dbe478
"Recraft": fc591c24
"Reve": 02eb39a9
"Rodin 3D": 2eb41834
"Runway": bd14d7ff
- "Sonilo": 48b9c581
+ "Sonilo": 66cc2806
"sync.so": 590402a9
"Tencent": cbe22aa6
- "Topaz": 5d08cf76
+ "Topaz": f557fb17
"Tripo": 842e861e
- "Pruna": f1b6a267
+ "Pruna": 08c63f88
"VEED": 1818bad3
"Vidu": be19d9ca
"Wan": 99df0ce3
"WaveSpeed": 349a7c9d
- "xAI": 28ff86de
+ "xAI": 6efd8650
"Cloud GPU": e228bb68
---
@@ -79,6 +81,7 @@ translationBlockHashes:
| :--------- | -----------------: | --------------------------: | --------------------------: | ---------------------: | :------------------ |
| Opus 5.5 | 1.207 | 1.509 | 2.414 | 0.121 | 6.035 |
| Opus 5、Opus 4.8 | 1.509 | 1.886 | 3.017 | 0.151 | 7.543 |
+| Fable 5.2 | 2.414 | 3.017 | 4.828 | 0.241 | 12.069 |
| Fable 5.1、Fable 5 | 3.017 | 3.772 | 6.035 | 0.302 | 15.087 |
| Sonnet 5.5、Sonnet 5 | 0.603 | 0.754 | 1.207 | 0.060 | 3.017 |
@@ -104,17 +107,19 @@ translationBlockHashes:
### 图像
-| 节点 | 积分 |
-| :------------------------- | :---------------------------: |
-| FLUX 1.1 [pro] Ultra 图像 | 12.66 / 次 |
-| FLUX.1 Kontext [pro] 图像 | 8.44 / 次 |
-| FLUX.1 Kontext [max] 图像 | 16.88 / 次 |
-| FLUX.1 展开图像 | 10.55 / 次 |
-| FLUX.1 Fill 图像 | 10.55 / 次 |
-| FLUX 擦除图像 | 6.33 / 次 + 0.84 / 额外 MP |
-| FLUX 虚拟试穿 | 7.91 / 次 + 1.06 / 额外 MP |
-| FLUX.2 [pro] | 6.33 / 次 + 3.17 / 额外 MP |
-| FLUX.2 [max] | 14.77 / 次 + 6.33 / 额外 MP |
+| 节点 | 积分 | 积分(含参考图) |
+| :------------------------ | :-------------------------: | :---------------------------------: |
+| FLUX 1.1 [pro] Ultra 图像 | 12.66 / 次 | — |
+| FLUX.1 Kontext [pro] 图像 | 8.44 / 次 | — |
+| FLUX.1 Kontext [max] 图像 | 16.88 / 次 | — |
+| FLUX.1 展开图像 | 10.55 / 次 | — |
+| FLUX.1 Fill 图像 | 10.55 / 次 | — |
+| FLUX 擦除图像 | 6.33 / 次 + 0.84 / 额外 MP | — |
+| FLUX 虚拟试穿 | 7.91 / 次 + 1.06 / 额外 MP | — |
+| FLUX.2 [pro] | 6.33 / 次 + 3.17 / 额外 MP | 9.50 – 31.65 / 次 + 3.17 / 额外 MP |
+| FLUX.2 [max] | 14.77 / 次 + 6.33 / 额外 MP | 21.10 – 65.41 / 次 + 6.33 / 额外 MP |
+
+连接参考图后按第三列计费。纯文本的 **FLUX.2 [pro]** 为 **6.33 积分**,每增加 1 MP 输出再加 **3.17 积分**(`ceil(W × H / 1 MP) − 1`)。**FLUX.2 [max]** 为 **14.77 积分**,每增加 1 MP 输出再加 **6.33 积分**。连接参考图时,按输入兆像素加价:每张参考图按 `ceil(W × H / 1 MP)` 向上取整后相加(最多 8 张),[pro] 每 1 MP 加 **3.17 积分**,[max] 每 1 MP 加 **6.33 积分**。第三列是输出为 1 MP、参考输入为 1–8 MP 时的区间(**9.50–31.65** / **21.10–65.41**),每增加 1 MP 输出仍按第二列的单价加收。[pro] 的 **3.17** 为四舍五入后的单价;公布的 **9.50–31.65** 区间是参考输入 1–8 MP 时的线上徽章端点。
### 图像:FLUX 3 Image
@@ -149,14 +154,14 @@ translationBlockHashes:
### 视频:FLUX 3 Max
-总积分 = **(积分/秒) × `duration`**(`duration` 为 5–20 秒)。`reasoning_level` 和 `generate_audio` 不影响价格。发布促销价有效期至 **2026 年 11 月 7 日(00:00 UTC)**,之后按标准价计费。
+总积分 = **(积分/秒) × `duration`**(`duration` 为 5–20 秒)。`reasoning_level` 和 `generate_audio` 不影响价格。720p 12.82、1080p 21.88 积分/秒的推出期价格已于 2026 年 11 月 7 日(00:00 UTC)结束,下表为标准价,自该时间起生效。
-| 节点 | 分辨率 | 促销积分/秒 | 标准积分/秒 |
-| :------------------------ | :---- | :----- | :----- |
-| FLUX 3 Max Text to Video | 720p | 12.82 | 51.29 |
-| FLUX 3 Max Text to Video | 1080p | 21.88 | 87.50 |
-| FLUX 3 Max Image to Video | 720p | 12.82 | 51.29 |
-| FLUX 3 Max Image to Video | 1080p | 21.88 | 87.50 |
+| 节点 | 分辨率 | 积分/秒 |
+| :------------------------ | :---- | :--- |
+| FLUX 3 Max Text to Video | 720p | 51.29 |
+| FLUX 3 Max Text to Video | 1080p | 87.50 |
+| FLUX 3 Max Image to Video | 720p | 51.29 |
+| FLUX 3 Max Image to Video | 1080p | 87.50 |
### 视频:FLUX 视频编辑
@@ -172,10 +177,10 @@ translationBlockHashes:
| 节点 | 模式 | 积分 / MP / 秒 |
| :------------------ | :-------- | :------------- |
-| FLUX 视频放大 | 创意 | 15.09 |
-| FLUX 视频放大 | 精确 | 10.56 |
+| FLUX 视频放大 | 创意 | 14.39 |
+| FLUX 视频放大 | 精确 | 10.07 |
-1080p 输出(2.07 MP)在「创意」模式下每秒 31.24 积分,「精确」模式为 21.86。4K 输出(8.29 MP)分别为 125.10 和 87.54。
+1080p 输出(2.07 MP)在「创意」模式下每秒 29.83 积分,「精确」模式为 20.88。4K 输出(8.29 MP)分别为 119.34 和 83.54。
## Bria
@@ -212,6 +217,18 @@ translationBlockHashes:
| Bria 移除视频背景(透明) | 10.55 / 秒 |
| Bria 视频橡皮擦 | 13.58 / 秒 |
+## Breeze
+
+Breeze 的语音节点均按 **9.05 积分 / 1K 计费字符** 计费,CJK 字符按两个字符计算。文本转语音和音色设计按你输入的文本计费。音色克隆、音色重混和音色本地化使用同一单价,但字符数要等运行结束后才知道:重混和本地化按生成的预览文本计费,克隆按厂商返回的计费字符数计费。因此这三项的节点 badge 会显示大约区间:**0.66–2.66**、**0.72–1.52**、**0.76–2.29** 积分。
+
+| 节点 | 积分 |
+| :---------------- | :------------- |
+| Breeze 文本转语音 | 9.05 / 1K 字符 |
+| Breeze 音色设计 | 9.05 / 1K 字符 |
+| Breeze 音色克隆 | 9.05 / 1K 字符 |
+| Breeze 音色重混 | 9.05 / 1K 字符 |
+| Breeze 音色本地化 | 9.05 / 1K 字符 |
+
## 字节跳动
### 大语言模型:ByteDance Seed
@@ -232,15 +249,17 @@ translationBlockHashes:
| 模型 | 积分 / 张 |
| :-------------------------------------------------- | :------------------- |
+| seedream 5.5 (`seedream-5-5-261028`) | 6.03 (1K) / 10.56 (2K) |
| seedream 5.0 pro (`seedream-5-0-pro-260628`) | 9.50 (1K) / 18.99 (2K) |
| seedream 5.0 flash (`seedream-5-0-flash-260915`) | 5.43 |
| seedream 5.0 lite (`seedream-5-0-260128`) | 7.39 |
| seedream-4-5-251128 | 8.44 |
| seedream-4-0-250828 | 6.33 |
+| Seedream 5.5 图层分离 | 6.03 (1K, 1.5K) / 10.56 (2K) |
| Seedream 5.0 Pro 图层分离 | 6.79 (1K, 1.5K) / 13.58 (2K) |
| Seedream 5.0 Flash 图层分离 | 5.43 |
-Seedream 5.0 Pro 使用基于分辨率的计费方式:**1K**(≤2.61 百万像素,**9.50 积分**)和 **2K**(>2.61 百万像素,**18.99 积分**)。Seedream 5.0 Flash 在所有分辨率下均按每张 **5.43 积分** 的固定费率计费。每次请求的第一张图片免费。超出第一张的每张额外图片(包括输出图片和输入/参考图片)加收 **0.63 积分**。
+Seedream 5.0 Pro 使用基于分辨率的计费方式:**1K**(≤2.61 百万像素,**9.50 积分**)和 **2K**(>2.61 百万像素,**18.99 积分**)。Seedream 5.0 Flash 在所有分辨率下均按每张 **5.43 积分** 的固定费率计费。每次请求的第一张图片免费。超出第一张的每张额外图片(包括输出图片和输入/参考图片)加收 **0.63 积分**。**Seedream 5.5** 在约 2.6 百万像素以内每张 **6.03 积分**,超过后 **10.56 积分**,每连接一张参考图再加 **2.41 积分**。**Seedream 5.5 图层分离** 在 1K/1.5K 下为 **6.03 积分**,在 2K 下为 **10.56 积分**,按本次运行的图片数计费。
### 图片:ByteDance Image(已弃用)
@@ -254,24 +273,45 @@ Seedream 5.0 Pro 使用基于分辨率的计费方式:**1K**(≤2.61 百
### 视频:Seedance 2.5
-基于 Token 计费。预计 Token 消耗 = **(输入视频时长 + 输出视频时长)× 输出视频宽度 × 输出视频高度 × 输出视频帧率 ÷ 1024**。仅输入文本、图片或音频时,输入视频时长为 0。实际用量以 API 返回为准。
+基于 Token 计费。ComfyUI 线上价格徽章估算 Token 的方式是 **floor((输出宽度 × 输出高度 ÷ 1024)×(24 × 计费时长 + 1))**,再乘以下表的「积分 / 1K Token」费率(其中的 `+ 1` 表示多计 1 帧)。计费时长为输出时长;若连接了参考视频,再加上输入视频时长。仅输入文本、图片或音频时,输入视频时长为 0。实际用量以 API 返回为准。
+
+| 模型 | 分辨率 | 视频输入 | 积分 / 1K Token |
+| :----------------- | :--------- | :------- | --------------: |
+| Seedance 2.5 | 480p, 720p | 否 | 3.229 |
+| Seedance 2.5 | 480p, 720p | 是 | 1.931 |
+| Seedance 2.5 | 1080p | 否 | 3.530 |
+| Seedance 2.5 | 1080p | 是 | 2.112 |
+| Seedance 2.5 | 4k | 否 | 1.418 |
+| Seedance 2.5 | 4k | 是 | 0.875 |
+| Seedance 2.5 Flash | 480p, 720p | 否 | 2.112 |
+| Seedance 2.5 Flash | 480p, 720p | 是 | 1.297 |
+
+**视频输入**表示已连接参考视频。文本、图片和音频参考均视为无视频输入。线上徽章除上述公式外,不再另设最低 Token 计费门槛。
+
+**Seedance 2.5 Flash** 只生成 480p 和 720p,最长 30 秒,两档费率相同。5 秒 720p 成片在无视频输入时约需 230 积分。
+
+相同时长下,4k 单帧的 Token 数量约为 1080p 的四倍,因此每 1K Token 单价更低,但总费用更高。
+
+参考视频至少需要 2 秒,输入总计不超过 30 秒。按上面的 token 公式,16:9 输出(24 fps)的预计费用:
-| 模型 | 分辨率 | 视频输入 | 积分 / 1K Token |
-| :----------- | :-------- | :------- | -------------: |
-| Seedance 2.5 | 480p, 720p | 否 | 3.229 |
-| Seedance 2.5 | 480p, 720p | 是 | 1.931 |
-| Seedance 2.5 | 1080p | 否 | 3.530 |
-| Seedance 2.5 | 1080p | 是 | 2.112 |
+| 分辨率 | 5 秒输出,无视频输入 | 5 秒输出 + 2 秒视频输入 | 5 秒输出 + 30 秒视频输入 |
+| :--------- | ------------------: | --------------------: | ---------------------: |
+| 480p | 156.38 | 130.64 | 650.12 |
+| 720p | 351.58 | 293.72 | 1461.63 |
+| 1080p | 865.00 | 722.82 | 3596.98 |
+| 4k | 1389.91 | 1197.81 | 5960.70 |
-**视频输入**表示已连接参考视频。文本、图片和音频参考均视为无视频输入。输入包含视频时适用最低 Token 消耗限制:若预计消耗低于最低值,按最低消耗计费。
+连接参考视频后使用较低的每 1K token 费率,所以 5 秒成片加 2 秒参考,有可能比无视频输入的 5 秒更便宜。线上徽章看不到输入视频的实际时长,因此在已连接参考视频、输出为 5 秒时,显示的是「仅按 5 秒输出(有视频输入费率)」到「5 秒输出 + 30 秒输入」的区间。
-参考视频至少需要 2 秒,输入总计不超过 30 秒。16:9 输出(24 fps)的预计费用:
+使用 **Seedance 2.5 Draft** 模型运行时,节点生成 480p 预览(按上面第一张表的 480p Token 费率计费)并返回 `draft_task_id`。**ByteDance Seedance 2.5 Draft to Final Video** 节点把该草稿渲染为 1080p 或 4k 成片;成片沿用草稿的提示词、参考素材、时长和音频设置。该节点的线上徽章按同一公式给出与时长无关的积分区间(下限为无参考视频的 4 秒 16:9 成片,上限为 30 秒成片加 30 秒参考视频)。上限在 1080p 为 21:9,在 4k 为 4:3 或 3:4:
+
+| 节点 | 分辨率 | 积分 |
+| :------------------------------------------ | :----- | ----------------: |
+| ByteDance Seedance 2.5 Draft to Final Video | 1080p | 693.43 – 6202.64 |
+| ByteDance Seedance 2.5 Draft to Final Video | 4k | 1114.23 – 10214.09 |
+
+4 秒 1080p、16:9、无参考视频的成片为 693.43 积分;30 秒 1080p、16:9 成片配 30 秒参考视频为 6163.19 积分。两个数字都包含徽章多计的那 1 帧。不同画幅会略有差异:1080p 上限 6202.64 对应 21:9,4k 上限 10214.09 对应 4:3 或 3:4。最终以 API 实际用量为准。
-| 分辨率 | 5 秒输出,无视频输入 | 5 秒输出 + 2–30 秒视频输入 |
-| :--------- | ------------------: | ------------------------: |
-| 480p | 155.1 | 166.9 – 649.3 |
-| 720p | 348.8 | 375.4 – 1459.8 |
-| 1080p | 857.8 | 923.9 – 3592.7 |
### 视频:Seedance 2.0
480p 和 720p 费率相同。Fast 和 Mini 不支持 1080p 或 4K。
@@ -311,7 +351,7 @@ Seedream 5.0 Pro 使用基于分辨率的计费方式:**1K**(≤2.61 百
| seedance-1-0-pro-fast-251015 | — | 211 |
- **已弃用**:`seedance-1-5-pro-251215` 将于 2026 年 11 月 11 日退役,届时 BytePlus 会关停该模型。ComfyUI 也将在同一天将其从 ByteDance 1.x 节点中移除,因此本页已删除该模型的价格行。请改用 `seedance-1-0-pro-250528`、`seedance-1-0-pro-fast-251015` 或 Seedance 2.x 模型。
+ **已退役**:BytePlus 已于 2026 年 11 月 11 日关停 `seedance-1-5-pro-251215`,因此本页已删除该模型的价格行。请改用 `seedance-1-0-pro-250528`、`seedance-1-0-pro-fast-251015` 或 Seedance 2.x 模型。
**已退役**:`seedance-1-0-lite-t2v-250428` 和 `seedance-1-0-lite-i2v-250428`,以及 **ByteDance 参考图片转视频** 节点,已退役并从 ComfyUI 中移除。以下价格仅作参考保留。
@@ -403,7 +443,7 @@ Seedream 5.0 Pro 使用基于分辨率的计费方式:**1K**(≤2.61 百
## FishAudio
-**限时优惠**:2026 年 10 月 13 日至 11 月 12 日,Drama 3 按约 **3.02** 积分 / 1K 字节计费。表格中 `drama-3` 一行为标准价格,自 2026 年 11 月 13 日起适用。
+Drama 3 在 2026 年 10 月 13 日至 11 月 12 日期间按上线费率约 **3.02** 积分 / 1K 字节计费。表格中 `drama-3` 一行为标准价格,自 2026 年 11 月 13 日起适用。
| 节点 | 模型 | 积分 |
@@ -446,6 +486,9 @@ Seedream 5.0 Pro 使用基于分辨率的计费方式:**1K**(≤2.61 百
| 模型 | 输入积分 / 1K | 输出积分 / 1K |
| :------------------------------------------------------ | -----------------: | :------------------ |
+| Gemini 4 Hydrogen (`gemini-4-hydrogen`) | 0.4526 | 2.263 |
+| Gemini 4 Argon (`gemini-4-argon`) | 1.20692 | 6.0346 |
+| Gemini 4 Carbon (`gemini-4-carbon`) | 1.20692 | 6.0346 |
| Gemini 3.8 Flash (`gemini-3.8-flash`) | 0.4526 | 2.263 |
| Gemini 3.7 Flash (`gemini-3.7-flash`) | 0.4526 | 2.263 |
| Gemini 3.5 Flash (`gemini-3.5-flash`) | 0.3165 | 1.899 |
@@ -477,6 +520,10 @@ Seedream 5.0 Pro 使用基于分辨率的计费方式:**1K**(≤2.61 百
| | 720p | 30.57 |
| | 1080p | 45.87 |
| | 4k | 91.76 |
+| Gemini Omni Flash 1.2 (`gemini-omni-1.2-flash`) | 360p | 6.98 |
+| | 720p | 20.97 |
+| | 1080p | 31.46 |
+| | 4k | 62.92 |
### 视频:Google Veo
@@ -532,23 +579,33 @@ Seedream 5.0 Pro 使用基于分辨率的计费方式:**1K**(≤2.61 百
| HeyGen Video Translate | 速度模式 | 4.07 / 秒 |
| HeyGen Text to Speech | N/A | ~0.20 / 秒 |
| HeyGen Video 1.0 Reference to Video | 480p,无参考素材 | 3.02 / 秒 |
-| HeyGen Video 1.0 Reference to Video | 768p,无参考素材 | 4.53 / 秒 |
-| HeyGen Video 1.0 Reference to Video | 2K,无参考素材 | 13.58 / 秒 |
-| HeyGen Video 1.0 Reference to Video | 480p,有参考素材 | 6.03 / 秒 |
-| HeyGen Video 1.0 Reference to Video | 768p,有参考素材 | 9.05 / 秒 |
-| HeyGen Video 1.0 Reference to Video | 2K,有参考素材 | 27.16 / 秒 |
-| HeyGen Video 1.0 Image to Video | 480p | 3.02 / 秒 |
-| HeyGen Video 1.0 Image to Video | 768p | 4.53 / 秒 |
-| HeyGen Video 1.0 Image to Video | 2K | 13.58 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 768p,无参考素材 | 5.43 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 1080p,无参考素材 | 12.07 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 2K,无参考素材 | 15.09 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 480p,有参考素材 | 10.56 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 768p,有参考素材 | 15.09 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 1080p 或 2K,有参考素材 | 27.16 / 秒 |
+| HeyGen Video 1.0 Reference to Video | 第 3 张起的每张额外参考图 | 6.03 / 张 |
+| HeyGen Video 1.0 Image to Video | 480p | 3.02 / 秒 |
+| HeyGen Video 1.0 Image to Video | 768p | 5.43 / 秒 |
+| HeyGen Video 1.0 Image to Video | 1080p | 12.07 / 秒 |
+| HeyGen Video 1.0 Image to Video | 2K | 15.09 / 秒 |
+| HeyGen Video 1.0 Image to Video | 超过 15 秒,480p | 13.58 / 秒 |
+| HeyGen Video 1.0 Image to Video | 超过 15 秒,768p | 22.63 / 秒 |
+| HeyGen Video 1.0 Image to Video | 超过 15 秒,1080p 或 2K | 31.68 / 秒 |
| HeyGen Voice 1.0 Instant Clone | N/A | 免费 |
| HeyGen Voice 1.0 Text to Speech | `heygen-voice-1` | 9.05 / 1K 字符 |
| HeyGen Voice 1.0 Text to Speech | `heygen-voice-1-turbo` | 2.26 / 1K 字符 |
-**文本转语音定价**:上线期费率(至 2026 年 10 月 31 日)约为 **4.53** 积分 / 1K 字符(`heygen-voice-1`)和 **1.21** 积分 / 1K 字符(`heygen-voice-1-turbo`)。以上各行为标准价格,自 2026 年 11 月 1 日起适用。
+**文本转语音定价**:上线期费率(至 2026 年 11 月 1 日 00:00 UTC)约为 **4.53** 积分 / 1K 字符(`heygen-voice-1`)和 **1.21** 积分 / 1K 字符(`heygen-voice-1-turbo`)。以上各行为标准价格(原价),自该时间起适用。
Avatar Video 按所选数字人形象的**类型**按秒计费(Digital Twin 数字分身、Photo Avatar 照片级、Studio Avatar 工作室级)。**自动**会根据形象自动选择引擎,因此费率横跨上表的完整范围;自定义 avatar ID 按其自身类型计费。
-HeyGen Video 1.0 的片段时长为 5 到 15 秒,按所选分辨率的费率计费。Reference to Video 在未连接参考图或参考视频时按无参考素材的费率计费;每连接一个参考视频,计费时长增加 5 秒。**2K** 输出需要 `16:9` 或 `9:16` 的宽高比;`auto` 仅在没有参考素材时可用于 `2K`。
+**限时优惠**:五折优惠价适用至 2026 年 12 月 1 日(00:00 UTC)。上表中的 Video 1.0 行为当前优惠价,长片段费率除外,长片段已是标准价。自 2026 年 12 月 1 日(00:00 UTC)起按标准价(原价)计费:无参考素材和 Image to Video(≤15 秒)为每秒 **6.03 / 10.86 / 24.14 / 30.17** 积分;有参考素材为每秒 **21.12 / 30.17 / 54.31** 积分;第 3 张起的每张额外参考图为 **12.07 积分**。长片段费率保持不变。
+
+HeyGen Video 1.0 的片段时长为 5 到 30 秒,按秒计费。超过 15 秒的片段在两个节点都没有参考素材时按长片段那几行计费。连接了参考素材时不能选择 15 秒以上。只要连接了任意一张图或一段视频,**Reference to Video** 就会改用有参考素材的费率,即使前两张图没有单张费用。前两张参考图免费;从第 3 张起每张加 **6.03 积分**(2026 年 12 月 1 日 00:00 UTC 起该附加费变为 **12.07 积分**)。参考视频最长 15 秒,按有参考素材的每秒费率计费。
+
+`auto_duration` 用提示词中台词的长度代替 `duration`:8 到 30 秒,或 8 到 15 秒再加参考素材的时长。**1080p** 和 **2K** 输出需要 `16:9` 或 `9:16` 的宽高比;`auto` 仅在没有参考素材时可用于这两个分辨率。
## Higgsfield
@@ -659,15 +716,15 @@ HeyGen Video 1.0 的片段时长为 5 到 15 秒,按所选分辨率的费率
## Ideogram
-| 节点 | turbo | 角色参考 | 渲染速度 | 积分 / 张 |
-| :---------- | :---- | :------- | :------- | :-------------- |
-| Ideogram V1 | 否 | — | — | 18.1 |
-| Ideogram V1 | 是 | — | — | 6.03 |
-| Ideogram V2 | 否 | — | — | 24.14 |
-| Ideogram V2 | 是 | — | — | 15.09 |
-| Ideogram V4 | — | — | TURBO | 9.05 |
-| Ideogram V4 | — | — | DEFAULT | 18.1 |
-| Ideogram V4 | — | — | QUALITY | 30.17 |
+
+ **已弃用**:ComfyUI 已移除 **Ideogram V1** 和 **Ideogram V2** 节点,因此本页不再列出它们的价格。请改用 **Ideogram V4**。
+
+
+| 节点 | 渲染速度 | 积分 / 张 |
+| :---------- | :------- | :-------: |
+| Ideogram V4 | TURBO | 9.05 |
+| Ideogram V4 | DEFAULT | 18.1 |
+| Ideogram V4 | QUALITY | 30.17 |
| 节点 | 质量 | 分辨率 | 积分 / 张 |
| :--------------- | :------- | :----- | :-------------- |
@@ -726,13 +783,41 @@ HeyGen Video 1.0 的片段时长为 5 到 15 秒,按所选分辨率的费率
| Krea 3 | 2K | high | 24.14 | 6.03 - 21.12 |
| Krea 3 | 2K | ultra | 78.45 | 6.03 - 21.12 |
-2K 输出时,每张参考图按参考图自身的像素大小增加 **6.03 到 21.12 积分**;1K 输出时参考图会被降采样到约 1 百万像素,因此每张固定加收 **6.03 积分**。`ultra` 不论 `resolution` 取值如何都固定输出 2K,并且至少要接入一张参考图。
+2K 输出时,每张参考图加收 **6.03–21.12 积分**(低档 **6.03**,高档 **21.12**,节点按参考图像素数在这两个值之间缩放)。1K 输出时参考图会被降采样到约 1 百万像素,因此每张固定加收 **6.03 积分**。`ultra` 不论 `resolution` 取值如何都固定输出 2K,并且至少要接入一张参考图。
## Kling
### 视频:积分 / 秒
-总积分 = **(积分/秒) × `时长`**,**Kling 3.0 Omni 编辑视频** 和 **Kling 动作控制** 除外(按输出秒数计费;编辑模式下时长跟随输入视频)。
+总积分 = **(积分/秒) × `时长`**,**Kling 3.0 Omni 编辑视频**、**Kling 4.0 编辑视频** 和 **Kling 动作控制** 除外(按输出秒数计费;编辑模式下时长跟随输入视频)。
+
+#### kling-4.0
+
+`kling-4.0` 生成 3 到 30 秒的 720p 或 1080p 视频,`kling-4.0-flash` 生成 3 到 20 秒的 720p 视频。
+
+| 节点 | 模型 | 分辨率 | 积分 / 秒 | 有参考视频时积分 / 秒 |
+| :------------------- | :-------------- | :----- | :-------- | :-------------------- |
+| Kling 4.0 文生视频 | kling-4.0 | 720p | 47.37 | — |
+| Kling 4.0 文生视频 | kling-4.0 | 1080p | 70.91 | — |
+| Kling 4.0 图生视频 | kling-4.0 | 720p | 47.37 | — |
+| Kling 4.0 图生视频 | kling-4.0 | 1080p | 70.91 | — |
+| Kling 4.0 参考生视频 | kling-4.0 | 720p | 47.37 | 28.36 |
+| Kling 4.0 参考生视频 | kling-4.0 | 1080p | 70.91 | 42.54 |
+| Kling 4.0 文生视频 | kling-4.0-flash | 720p | 28.36 | — |
+| Kling 4.0 图生视频 | kling-4.0-flash | 720p | 28.36 | — |
+| Kling 4.0 参考生视频 | kling-4.0-flash | 720p | 28.36 | 19.01 |
+
+**Kling 4.0 参考生视频** 连接参考视频后,整段运行都按较低的费率计费:输出的每一秒和参考视频的每一秒都按该费率计算。单个片段为 3 到 30 秒(`kling-4.0-flash` 为 3 到 20 秒),所有片段合计不超过同一上限。5 秒的 720p 成片配 30 秒参考视频约 992.7 积分。
+
+**Kling 4.0 编辑视频** 一律按较低的(参考视频)费率计费:**积分 = 单价 ×(输入视频秒数 + 参考视频秒数)**。输出与输入等长,因此没有参考视频时,按输入视频每秒 **2 × 单价** 计费。有参考视频时,实际费用仍按同一公式;线上徽章看不到视频实际时长,所以显示的是输入视频每秒 **1 倍到 2 倍单价** 的区间。参考图不加单张费用。
+
+| 模型 | 分辨率 | 单价(积分 / 秒) | 无参考视频 | 有参考视频 |
+| :-------------- | :----- | ----------------: | :--------- | :--------- |
+| kling-4.0 | 720p | 28.36 | 56.73 / 秒输入(2 × 单价) | 28.36 ×(输入 + 参考秒数) |
+| kling-4.0 | 1080p | 42.54 | 85.09 / 秒输入(2 × 单价) | 42.54 ×(输入 + 参考秒数) |
+| kling-4.0-flash | 720p | 19.01 | 38.02 / 秒输入(2 × 单价) | 19.01 ×(输入 + 参考秒数) |
+
+10 秒的 1080p 编辑(无参考视频)约 850.9 积分。
#### kling-video-o1
@@ -788,10 +873,6 @@ HeyGen Video 1.0 的片段时长为 5 到 15 秒,按所选分辨率的费率
| 动作控制 | — | 标准 | 14.77 |
| 动作控制 | — | 专业 | 23.63 |
-
- **已退役(EOL)**:Kling V1.5、V1.6、V2.1、V2.1 Master 及 Kolors 虚拟试穿计划于 **2026 年 9 月 15 日** 退役。以下定价保留供参考,直至移除。
-
-
#### kling-v2-5-turbo
**Kling 文生视频**、**Kling 图生视频(首帧)** 和 **Kling 首尾帧生视频** 仅使用 `kling-v2-5-turbo` 的 **专业** 模式,按固定 5 秒或 10 秒时长按次计费。
@@ -839,6 +920,10 @@ HeyGen Video 1.0 的片段时长为 5 到 15 秒,按所选分辨率的费率
| :------------------- | :-------- |
| Kling Virtual Try-On | 33.79 |
+
+ **已退役**:Kling 已于 **2026 年 9 月 15 日** 退役 V1.5、V1.6、V2.1、V2.1 Master,以及早前的 Kolors 虚拟试穿服务。Kling 节点不再提供这些模型,本页也因此移除了它们的定价行。上方的 **Kling Virtual Try-On** 行是另一个仍在提供的节点,按次计费。
+
+
## Lightricks
### 视频:积分 / 秒
@@ -868,15 +953,35 @@ HeyGen Video 1.0 的片段时长为 5 到 15 秒,按所选分辨率的费率
分辨率映射到 `1280x720` / `720x1280`(720p)、`1920x1080` / `1080x1920`(1080p)、`2560x1440` / `1440x2560`(1440p)和 `3840x2160` / `2160x3840`(4K)。**LTX-2.5 (Fast)** 超过 10 秒的时长需要 720p 或 1080p 分辨率和 24/25 FPS;**LTX-2 (Fast)** 超过 10 秒的时长仅 1080p 和 25 FPS 下可用。
+### 后处理:积分 / 秒
+
+这些节点处理输入的视频,按输入视频的秒数计费。
+
+| 节点 | 分辨率 | 积分 / 秒 |
+| :------------------- | ------ | --------------- |
+| LTX Video AlphaGen | 1080p(短边 ≤1080,低于 1080 按此档) | 30.17 |
+| LTX Video AlphaGen | 1440p(短边 ≤1440) | 75.43 |
+| LTX Video AlphaGen | 2160p(短边 ≤2160) | 301.73 |
+| LTX Video SDR to HDR | 1080p(≤ 1920×1080) | 120.69 |
+| LTX Video SDR to HDR | 1440p(≤ 2560×1440) | 241.39 |
+| LTX Video SDR to HDR | 4K(≤ 3840×2160) | 482.77 |
+| LTX Video Reframe | 720p | 30.17 |
+| LTX Video Reframe | 1080p | 60.35 |
+| LTX Video Upscale | 1080p | 24.14 |
+| LTX Video Upscale | 4k | 90.52 |
+| LTX Video Upscale | 8k | 362.08 |
+
+**AlphaGen** 按输入视频每秒计费,档位由输入画面的短边(宽和高中较小的一边)决定;短边低于 1080 像素时按 1080p 计费。**SDR to HDR** 也按输入视频每秒计费,并取能覆盖该输入像素数的最低分辨率档。线上徽章看不到输入分辨率,因此只显示区间(AlphaGen 为每秒 **30.17–301.73** 积分,SDR to HDR 为每秒 **120.69–482.77** 积分)。实际扣费按表中的三档费率计算。**AlphaGen** 最多接受 10 秒视频,**Reframe** 最多 60 秒;**SDR to HDR** 和 **Upscale** 按分辨率限制帧数(SDR to HDR 在 1080p 下 273 帧,Upscale 为 2,401 帧)。
+
## Luma
### 视频
- **已弃用**:旧版 **Luma 文本转视频** 和 **Luma 图像转视频** 节点已弃用,Luma 已于 2026 年 10 月 24 日移除其使用的接口,因此这两个节点已无法生成视频。请改用 **Luma Ray 3.2** 节点。下表中的 ray-2、ray-flash-2 和 ray-1-6 行仅作参考保留;**ray-1-6** 自 2026 年 9 月起已停止提供。
+ **已退役**:旧版 Luma 文本转视频、图像转视频和图像生成节点已移除,ray-2、ray-flash-2、ray-1-6、photon-1 和 photon-flash-1 模型也不再提供。下表只列出 ComfyUI 当前仍提供的节点。
-固定 **/次** 价格与每个 `duration` 选项的节点徽章一致。**编辑**和**重构图**按 **(积分/秒) × 视频长度**计费。**ray-1-6** 始终 **105.50 / 次**。
+固定 **/次** 价格与每个 `duration` 选项的节点徽章一致。**编辑**和**重构图**按 **(积分/秒) × 视频长度**计费。
| 模型 | 任务 | 分辨率 | 积分 |
| :---------- | :--------- | :--------- | :------------------------------------------------ |
@@ -892,24 +997,13 @@ HeyGen Video 1.0 的片段时长为 5 到 15 秒,按所选分辨率的费率
| ray-3.2 | 重构图 | 540p | 12.66 / 秒 |
| ray-3.2 | 重构图 | 720p | 25.32 / 秒 |
| ray-3.2 | 重构图 | 1080p | 75.96 / 秒 |
-| ray-2 | 生成 | 540p | 120.27 / 次 (5s), 217.33 / 次 (9s) |
-| ray-2 | 生成 | 720p | 215.22 / 次 (5s), 386.13 / 次 (9s) |
-| ray-2 | 生成 | 1080p | 478.97 / 次 (5s), 865.10 / 次 (9s) |
-| ray-2 | 生成 | 4K | 1922.21 / 次 (5s), 3460.40 / 次 (9s) |
-| ray-flash-2 | 生成 | 540p | 42.20 / 次 (5s), 75.96 / 次 (9s) |
-| ray-flash-2 | 生成 | 720p | 71.74 / 次 (5s), 128.71 / 次 (9s) |
-| ray-flash-2 | 生成 | 1080p | 166.69 / 次 (5s), 299.62 / 次 (9s) |
-| ray-flash-2 | 生成 | 4K | 660.43 / 次 (5s), 1192.15 / 次 (9s) |
-| ray-1-6 | 生成 | — | 105.50 / 次 |
-
-**Luma Ray 3.2** 图片转视频和视频延长始终按 **5s** 价格计费。旧版图片转视频使用 **ray-2** / **ray-flash-2** 行中的 **5s** 或 **9s** 价格。
+
+**Luma Ray 3.2** 图片转视频和视频延长始终按 **5s** 价格计费。
### 图片
| 模型 | 任务 | 积分 |
| :------------- | :--------- | :----------------------------------- |
-| photon-flash-1 | — | 0.57 / 次 |
-| photon-1 | — | 2.19 / 次 |
| uni-1 | 生成 | 8.52 / 次 + 0.63 / 参考图片 (≤ 9) |
| uni-1 | 编辑 | 9.16 / 次 + 0.63 / 参考图片 (≤ 8) |
| uni-1-max | 生成 | 21.10 / 次 + 0.63 / 参考图片 (≤ 9) |
@@ -975,7 +1069,7 @@ Ultra 模式(适用于 Meshy 7.1、Meshy 7 和 `latest` 模型)增加 60.3
| MiniMax H3 首尾帧转视频 | MiniMax H3 | 2K | 39.22 / 秒 |
| MiniMax H3 参考转视频 | MiniMax H3 | 768P | 27.16 / 秒,超过 5 张图片每张加 12.07 |
| MiniMax H3 参考转视频 | MiniMax H3 | 2K | 39.22 / 秒,超过 5 张图片每张加 12.07 |
-| MiniMax H3 上下文增强 | MiniMax H3 | — | 4.22 / 次(仅文本),10.55 / 次(含首尾帧),12.66–23.21 / 次(含参考媒体) |
+| MiniMax H3 上下文增强 | MiniMax H3 | — | 271.56 / 100 万输入 token + 1,086.23 / 100 万输出 token |
| MiniMax H3 重新生成至 2K | MiniMax H3 | — | 15.09 / 秒 |
| MiniMax H3 文本转视频 | MiniMax H3 Max | 480P | 15.09 / 秒 |
| MiniMax H3 文本转视频 | MiniMax H3 Max | 768P | 24.14 / 秒 |
@@ -1008,7 +1102,7 @@ Ultra 模式(适用于 Meshy 7.1、Meshy 7 和 `latest` 模型)增加 60.3
**MiniMax Hailuo 02 Video** 在 1080P 下仅支持 **6 秒**。
-**MiniMax H3** 按秒计费(时长 4–15 秒)。**MiniMax H3 参考转视频** 超过 5 张参考图片每张加 12.07 积分,参考视频按 27.16 积分/秒(768P)或 39.22 积分/秒(2K)计费(每段视频 2 秒起,总计最多 15 秒,近似值)。**MiniMax H3 上下文增强** 的价格取决于附加媒体:仅文本、首尾帧或参考媒体(近似值)。
+**MiniMax H3** 按秒计费(时长 4–15 秒)。**MiniMax H3 参考转视频** 超过 5 张参考图片每张加 12.07 积分,参考视频按 27.16 积分/秒(768P)或 39.22 积分/秒(2K)计费(每段视频 2 秒起,总计最多 15 秒,近似值)。**MiniMax H3 上下文增强** 按 **每 100 万输入 token 271.56 积分**、**每 100 万输出 token 1,086.23 积分** 计费。节点徽章在构图时看不到 token 数,因此显示近似区间:约 **4.22**(仅文本)、**10.55**(含首尾帧)或 **12.66–23.21**(含参考媒体)。
**MiniMax H3 Max** 按秒计费,费率见上表。**文本转视频**、**首尾帧转视频**和**参考转视频**的输出时长为 **5–15 秒**。**对口型** 可接受 5 至约 899 秒音频;输出超过 15 秒时适用下文 **1.2×** 倍率。
@@ -1034,6 +1128,19 @@ Ultra 模式(适用于 Meshy 7.1、Meshy 7 和 `latest` 模型)增加 60.3
**MiniMax H3 Max 人物替换**、**视频延长**、**重新构图**、**视频插入**、**对口型** 和 **镜头控制** 均按视频秒数计费。视频延长还可选 Turbo 模型,费率为 Max 的一半。**重新构图**按表中费率按秒计费。较长的源视频使用重叠窗口,因此计费秒数可能超过最终输出时长。**对口型**按表中费率按秒计费。输出**超过 15 秒**时,适用 **1.2×** 倍率(例如 768P → 28.97 / 秒)。视频插入在免费 4,096 Token 额度之外按每 1K Token 6.03 积分计费,与参考转视频共用同一额度:基础 240 Token,每张图片 2,048 Token,每段参考视频在 768P 输出下 225,000 Token(480P 为 87,000)。
+### 图像:MiniMax H3 Image
+
+每张生成的图片计费。费用由输出尺寸,以及按 Token 计费的提示词和参考图共同决定。
+
+| 节点 | 模型 | 尺寸 | 积分 / 张 |
+| :--------------- | -------------------------- | ---- | ------------- |
+| MiniMax H3 Image | MiniMax-H3-Image | 2K | 5.32 – 8.17 |
+| MiniMax H3 Image | MiniMax-H3-Image | 4K | 19.41 – 22.26 |
+| MiniMax H3 Image | MiniMax-H3-Image-Highspeed | 2K | 10.02 – 12.86 |
+| MiniMax H3 Image | MiniMax-H3-Image-Highspeed | 4K | 38.20 – 41.04 |
+
+上表是纯文本且开启 `enhance_prompt` 时的结果。完整公式为:输出积分 = `max(像素数, 4,194,304) / 1,024 × 0.00114657`(Highspeed 用 `0.00229315`),即 2K(4 MP)为 **4.70 积分**,4K 约 **18.79 积分**,Highspeed 为 **9.39** / **37.57**。提示词 Token 按每个 **0.000694 积分** 计费:无参考图且关闭增强为 **0–300**,无参考图且开启增强为 **900–5,000**,有 `n` 张参考图时为 **2,000 + 300 × (n − 1)** 到 **6,500 + 800 × (n − 1)**。每张参考图另外加上 **2,050–8,194** 个图像 Token,每个 **0.000453 积分**(Highspeed 为 **0.000905**),约 **0.93 到 3.71 积分**(Highspeed 为 **1.86 到 7.42 积分**)。关闭增强时,2K 每张约 **4.70 积分**。自定义尺寸按自身面积计费,最低按 4 百万像素计算。
+
## OpenAI
DALL·E 和 Sora 使用固定的每张图片或每秒费率。**OpenAI GPT Image 2** 和 **OpenAI ChatGPT** 按 API Token 用量计费。
@@ -1084,14 +1191,55 @@ DALL·E 和 Sora 使用固定的每张图片或每秒费率。**OpenAI GPT Image
输出令牌数随尺寸和质量变化(例如 1024x1024 `low` 约为 196 个输出令牌,按 `gpt-image-1.5` 费率约 1.32 积分)。每张 1024x1024 参考图约增加 2.47 积分的图像输入,提示词文本按文本输入费率计费。生成数量(`n`)会等比放大总费用。
+**GPT Image 2.6** 模型(`gpt-image-2.6-flare`、`gpt-image-2.6-sunburst`)共用一套费率。下表为 1024x1024 每张图片的积分:
+
+| 质量 | 积分 / 张 |
+| :----- | :-------- |
+| low | 7.45 |
+| medium | 13.12 |
+| high | 15.89 |
+| xhigh | 28.25 |
+| max | 63.57 |
+
+尺寸越大费用越高:3840x2160 在 `low` 下为 14.12 积分,在 `max` 下为 120.8 积分。总积分 = **(输出价格 + 参考图数量 × 2.47 到 3.71) × `n`**。每张参考图另加 **2.47–3.71 积分**。线上徽章看不到决定该单价的参考图尺寸,因此显示这个区间;生成数量(`n`)会等比放大总费用。
+
+### 图片:GPT Image 2.6 GIF
+
+**OpenAI GPT Image 2.6 GIF** 按次计费,不按 Token 计费。每一帧都是一张完整图片,因此画面尺寸、质量和 `frame_count` 决定价格。`frame_count` 按 2 到 8 帧、9 帧、10 到 12 帧、13 到 16 帧四档计费,尺寸越大能选的档位越少。下表为 512x512、9 帧的积分:
+
+| 质量 | 积分 / 次 |
+| :----- | :-------- |
+| low | 10.61 |
+| medium | 18.76 |
+| high | 22.72 |
+| xhigh | 40.41 |
+| max | 90.92 |
+
+384x384 跑 2 到 8 帧在 `low` 下为 3.88 积分,720x720 跑 13 到 16 帧在 `max` 下为 214.69 积分。`frame_duration_ms` 不影响价格。
+
+### 图片:GPT Image 2.6 SVG
+
+**OpenAI GPT Image 2.6 SVG** 按所选尺寸和品质收取静图价格,再加上 **0.23–1.77 积分** 的矢量处理费用。线上徽章看不到决定该矢量处理费用的输入,因此在静图价格之上显示这个区间;下表区间就是静图价格加上该区间。SVG 徽章不再另行加收参考图费用。尺寸可选 1024x1024、1024x1536 和 1536x1024。下表为 1024x1024 的积分:
+
+| 质量 | 积分 / 次 |
+| :----- | :----------- |
+| low | 7.68 – 9.22 |
+| medium | 13.36 – 14.90 |
+| high | 16.12 – 17.66 |
+| xhigh | 28.48 – 30.03 |
+| max | 63.81 – 65.35 |
+
+1024x1536 或 1536x1024 更便宜:`low` 下为 6.25 到 7.79 积分,`max` 下为 49.90 到 51.44 积分。
+
### 聊天
基于 Token 计费。徽章费率为 **每 1K Token**;表格为 **每 1M Token**。
| 模型 | 输入积分 / 1M | 输出积分 / 1M |
| :----------- | -----------------: | :------------------ |
+| gpt-6.1-sol、gpt-6-sol | 603.46 | 3017.3 |
+| gpt-6.1-terra | 301.73 | 1508.65 |
| gpt-6-astra | 3017.3 | 15086.5 |
-| gpt-6-sol | 603.46 | 3017.3 |
| gpt-6-luna | 30.17 | 150.87 |
| gpt-5.6-sol、gpt-5.6-sol-pro | 1055 | 6330 |
| gpt-5.6-terra、gpt-5.6-terra-pro | 527.5 | 3165 |
@@ -1151,10 +1299,12 @@ DALL·E 和 Sora 使用固定的每张图片或每秒费率。**OpenAI GPT Image
| x-ai/grok-4.5 | 603.46 | 1810.38 |
| x-ai/grok-4.20 | 377.16 | 754.33 |
| x-ai/grok-4.3 | 377.16 | 754.33 |
+| meta/muse-wonder | 1206.92 | 6034.60 |
+| meta/muse-spark-1.3 | 377.16 | 1282.35 |
+| meta/muse-ember | 30.17 | 150.87 |
| deepseek/deepseek-v4-pro | 131.25 | 262.51 |
| deepseek/deepseek-v4-flash | 33.79 | 67.59 |
| deepseek/deepseek-v3.2 | 76.04 | 114.05 |
-| qwen/qwen3.6-max-preview | 313.80 | 1882.80 |
| qwen/qwen3.6-plus | 98.06 | 588.37 |
| qwen/qwen3.6-flash | 56.57 | 339.45 |
| mistralai/mistral-large-2512 | 150.87 | 452.60 |
@@ -1196,6 +1346,39 @@ DALL·E 和 Sora 使用固定的每张图片或每秒费率。**OpenAI GPT Image
价格为近似值。1 MP 材质只启用默认的 `basecolor` 与 `normal` 贴图时为 15.09 积分;4 MP、五张贴图全部启用且 4× 放大时,**Patina Material** 为 184.06 积分,**Patina Material Extract** 为 211.21 积分。
+## Pika
+
+Pika H3 Flex 按视频秒数计费。总积分 = **(积分 / 秒)× 计费秒数**。`duration` 为 5 到 15 秒。`480p` 与 `768p` 的计费时长等于 `duration`。从 `1080p` 起,节点会把片段对齐到帧网格,因此 5 秒的生成按 5.17 秒计费,10 秒按 10.13 秒计费,15 秒按 15.08 秒计费。
+
+| 节点 | 分辨率 | 积分 / 秒 |
+| :------------------------------ | :----- | :-------- |
+| Pika H3 Flex Video | 480p | 3.02 |
+| Pika H3 Flex Video | 768p | 6.03 |
+| Pika H3 Flex Video | 1080p | 10.56 |
+| Pika H3 Flex Video | 2k | 12.07 |
+| Pika H3 Flex Video | 4k | 24.14 |
+| Pika H3 Flex Reference to Video | 480p | 9.05 |
+| Pika H3 Flex Reference to Video | 768p | 18.10 |
+| Pika H3 Flex Reference to Video | 1080p | 22.63 |
+| Pika H3 Flex Reference to Video | 2k | 24.14 |
+| Pika H3 Flex Reference to Video | 4k | 30.17 |
+
+**Pika H3 Flex Video** 以 480p 生成 5 秒需要 15.09 积分,以 1080p 生成 5 秒需要 54.56 积分。
+
+**Pika H3 Flex Reference to Video** 会用同一费率对连接的参考视频计费。单个片段为 2 到 15 秒,参考视频总时长上限为 15 秒,因此费用取决于你接入多少参考视频。480p 下生成 5 秒并带一个 2 秒参考视频需要 63.36 积分。4k 下生成 5 秒并带 15 秒参考视频需要 608.49 积分。以上 Pika 算例均按未取整的每秒费率计算。
+
+### 音频:Pika
+
+Pika 的音频节点同样按秒计费。**Pika Soundtrack** 按输入视频时长计费(最长 10 分钟),**Pika SFX** 按 `duration` 计费(1 到 20 秒),**Pika Music** 按曲目时长计费:`text to music` 用 `duration`(10 到 360 秒),`cover` 用源音频时长并限制在 5 到 360 秒。
+
+| 节点 | 积分 / 秒 |
+| :-------------- | :-------- |
+| Pika Soundtrack | 1.509 |
+| Pika SFX | 0.060 |
+| Pika Music | 0.075 |
+
+10 分钟的 Soundtrack 需要 905.2 积分,20 秒的 SFX 需要 1.21 积分,3 分钟的 Music 曲目需要 13.58 积分。
+
## Pixverse
模型 **v3.5**(API 中固定,不可选择)。
@@ -1222,19 +1405,12 @@ PixVerse V6 节点可生成 1 到 15 秒的视频,并可选生成原生音频
| PixVerse V6 | 540p | 7.18 | 9.23 |
| PixVerse V6 | 720p | 9.23 | 12.31 |
| PixVerse V6 | 1080p | 18.47 | 23.60 |
-| PixVerse V6 Max | 360p | 6.71 | 9.39 |
-| PixVerse V6 Max | 540p | 9.39 | 12.07 |
-| PixVerse V6 Max | 720p | 12.07 | 16.09 |
-| PixVerse V6 Max | 1080p | 24.14 | 30.84 |
+| PixVerse V6 Max | 360p | 8.05 | 12.07 |
+| PixVerse V6 Max | 540p | 12.07 | 14.75 |
+| PixVerse V6 Max | 720p | 14.75 | 20.12 |
+| PixVerse V6 Max | 1080p | 29.50 | 37.55 |
-**PixVerse V6 Max** 在 **2026 年 11 月 3 日**之前按上表的优惠费率计费。自该日期起,每秒费率如下:
-
-| 质量 | 无音频 | 带音频 |
-| :---- | :----- | :----- |
-| 360p | 8.05 | 12.07 |
-| 540p | 12.07 | 14.75 |
-| 720p | 14.75 | 20.12 |
-| 1080p | 29.50 | 37.55 |
+**PixVerse V6 Max** 的优惠费率已于 **2026 年 11 月 2 日**结束,上表各行为 **2026 年 11 月 3 日**起生效的标准费率。
**PixVerse V6 Fusion** 按参考视频每秒的费率 2 倍计费,而非输出时长。Fusion 仅使用 `PixVerse V6` 的费率。
@@ -1242,21 +1418,10 @@ PixVerse V6 节点可生成 1 到 15 秒的视频,并可选生成原生音频
### 图片
-已停用的 **arrow-1.1** 系列按次收取固定费用:
-
- **已弃用**:Quiver 已于 **2026 年 10 月 16 日** 停用 Arrow 1 系列模型,Text to SVG 和 Image to SVG 节点中不再提供 `arrow-1.1`、`arrow-1.1-max` 和 `arrow-preview`。请改用 Arrow 2 模型。以下价格仅作参考保留。
+ **已停用**:Quiver 已于 **2026 年 10 月 16 日** 停用 Arrow 1 系列模型(`arrow-1.1`、`arrow-1.1-max`、`arrow-preview`),Text to SVG 和 Image to SVG 节点中不再提供这些模型。请改用下面的 Arrow 2 模型。
-| 模型 | 节点 | 积分 / 次 |
-| :------------ | :------------------ | :---------------- |
-| arrow-1.1 | Quiver Text to SVG | 60.35 |
-| arrow-1.1 | Quiver Image to SVG | 45.26 |
-| arrow-1.1-max | Quiver Text to SVG | 75.43 |
-| arrow-1.1-max | Quiver Image to SVG | 60.35 |
-| arrow-preview | Quiver Text to SVG | 90.52 |
-| arrow-preview | Quiver Image to SVG | 90.52 |
-
**arrow-2** 与 **arrow-2-telos** 按请求消耗的 Token 计费:
| 模型 | 输入积分 / 1M Token | 输出积分 / 1M Token |
@@ -1377,12 +1542,15 @@ Gen4 Turbo:**5s** 或 **10s**。总积分 = **(积分/秒) × `时长`**。
## Sonilo
-文本转音乐:总积分 = **(积分/秒) × `时长`**(1–360 秒)。
+视频转音乐按输入视频每秒计费,最长 6 分钟。文本转音乐与 Music 1.5 按音频每秒计费:总积分 = **(积分/秒) × `时长`**。
| 节点 | 积分 |
| :-------------------- | :----------- |
-| Sonilo Video to Music | 1.9 / 秒 |
-| Sonilo Text to Music | 0.5275 / 秒 |
+| Sonilo Video to Music | 2.72 / 秒 |
+| Sonilo Text to Music | 0.6789 / 秒 |
+| Sonilo Music 1.5 | 0.6789 / 秒 |
+
+文本转音乐的时长范围为 1 到 360 秒。Music 1.5 可用 `auto` 按歌词确定长度,也可用 `custom` 指定 10 到 360 秒。
## sync.so
@@ -1453,7 +1621,7 @@ Gen4 Turbo:**5s** 或 **10s**。总积分 = **(积分/秒) × `时长`**。
| :-------- | :------------- |
| Reimagine | 67.52 / 8MP |
| Bloom 2 | 96.55 / 8MP |
-| Wonder 3.5 | 24.14 / 8MP |
+| Wonder 3.5, Wonder 4 | 24.14 / 8MP |
**Topaz Image Enhance(旧版)** 已弃用。它没有固定的运行前估算;费用来自轮询时 API 的估算(API 积分 × 0.08 USD)。按输入分辨率的典型费用:
@@ -1614,6 +1782,19 @@ Pruna P-Video-2 按生成视频的秒数计费。总积分 = **(积分 / 秒) ×
| | 1080p | 关 | 15.09 |
| | 1080p | 开 | 9.05 |
+### 视频:P-Video Upscale
+
+Pruna P-Video Upscale 按输出视频的秒数计费。总积分 = **(积分 / 秒) × 输出时长**。费率随 `target` 变化,输出高于 30 fps 时按两倍计费,源视频大于目标尺寸时保持原尺寸。
+
+`720p`、`1080p`、`2K`、`4K` 的上线费率为每秒 1.51、3.02、6.03、12.07 积分,适用于 2026 年 11 月 12 日(17:00 UTC)之前。下表为标准费率,自该日起生效。
+
+| target | 目标尺寸 | 积分 / 秒 |
+| :----- | :-------- | :-------- |
+| 720p | 约 0.9 MP | 3.02 |
+| 1080p | 约 2.1 MP | 6.03 |
+| 2K | 约 3.7 MP | 12.07 |
+| 4K | 约 8.3 MP | 24.14 |
+
## VEED
| 节点 | 选项 | 积分 |
@@ -1703,23 +1884,25 @@ Pruna P-Video-2 按生成视频的秒数计费。总积分 = **(积分 / 秒) ×
| wan2.5-t2i-preview, wan2.5-i2i-preview | — | 6.33 / 次 |
| wan2.7-t2v, wan2.7-i2v | 720P | 21.1 / 秒 |
| | 1080P | 31.65 / 秒 |
-| wan2.7-i2v (续作) | 720P | 21.1 / 秒 (范围) |
-| | 1080P | 31.65 / 秒 (范围) |
+| wan2.7-i2v (续作) | 720P | 21.1 / 秒(输出 + 2–5 秒输入) |
+| | 1080P | 31.65 / 秒(输出 + 2–5 秒输入) |
| wan2.7-videoedit | 720P | 21.1 / 秒 (输入+输出) |
| | 1080P | 31.65 / 秒 (输入+输出) |
-| wan2.7-r2v | 720P | 21.1 / 秒 (范围) |
-| | 1080P | 31.65 / 秒 (范围) |
-| wan3.0-video | 480P | 15.09 / 秒 (范围) |
-| | 720P | 30.17 / 秒 (范围) |
-| | 1080P | 60.35 / 秒 (范围) |
-| wan3.0-video-prime | 480P | 20.52 / 秒 (范围) |
-| | 720P | 42.24 / 秒 (范围) |
-| | 1080P | 84.48 / 秒 (范围) |
+| wan2.7-r2v | 720P | 21.1 / 秒(输出至 输出 + 5 秒输入) |
+| | 1080P | 31.65 / 秒(输出至 输出 + 5 秒输入) |
+| wan3.0-video | 480P | 15.09 / 秒 |
+| | 720P | 30.17 / 秒 |
+| | 1080P | 60.35 / 秒 |
+| wan3.0-video-prime | 480P | 20.52 / 秒 |
+| | 720P | 42.24 / 秒 |
+| | 1080P | 84.48 / 秒 |
| wan2.5-t2v-preview, wan2.5-i2v-preview | 480p | 10.55 / 秒 |
| wan2.5-t2v-preview, wan2.5-i2v-preview, wan2.6-t2v, wan2.6-i2v | 720p | 21.1 / 秒 |
| wan2.5-t2v-preview, wan2.5-i2v-preview, wan2.6-t2v, wan2.6-i2v | 1080p | 31.65 / 秒 |
-| wan2.6-r2v | 720p | 21.1 / 秒 (范围) |
-| | 1080p | 31.65 / 秒 (范围) |
+| wan2.6-r2v | 720p | 21.1 / 秒(输出 + 2–5 秒输入) |
+| | 1080p | 31.65 / 秒(输出 + 2–5 秒输入) |
+
+**Wan 2.6 参考转视频** 和 **Wan 2.7 续作** 按输出时长加上 **2–5 秒** 输入计费,单价与表中每秒费率相同。**Wan 2.7 参考转视频** 的范围是仅输出时长,到输出时长再加 **5 秒** 输入。**Wan 2.7 Edit** 按表中费率对输入加输出计费。**Wan 3.0** 按表中每秒费率计费:若 `duration` 为 `auto`,徽章显示该每秒单价;否则范围为所选时长到 `min(duration + 15, 30)` 秒。
## WaveSpeed
@@ -1744,17 +1927,23 @@ Pruna P-Video-2 按生成视频的秒数计费。总积分 = **(积分 / 秒) ×
| :------------------------- | :--------- | :------- | :------------------------------------------- |
| grok-imagine-image-2.0 | 1K | 生成 | 12.66(中等)/ 8.44(低)每张输出图片 |
| | 2K | 生成 | 16.88(中等)/ 12.66(低)每张输出图片 |
-| grok-imagine-image-quality | 1K | 生成 | 10.55 / 输出图片 |
-| | 2K | 生成 | 14.77 / 输出图片 |
-| grok-imagine-image-pro | — | 生成 | 14.77 / 输出图片 |
+| grok-imagine-image-quality | 1K | 生成 | 8.44 / 输出图片 |
+| | 2K | 生成 | 12.66 / 输出图片 |
+| grok-imagine-image-pro | 1K | 生成 | 8.44 / 输出图片 |
+| | 2K | 生成 | 12.66 / 输出图片 |
| grok-imagine-image | — | 生成 | 4.22 / 输出图片 |
-| grok-imagine-image-2.0 | 1K | 编辑 | 2.11 输入 + 12.66(中等)/ 8.44(低)每张输出图片 |
-| | 2K | 编辑 | 2.11 输入 + 16.88(中等)/ 12.66(低)每张输出图片 |
-| grok-imagine-image-quality | 1K | 编辑 | 2.11 输入 + 10.55 / 输出图片 |
-| | 2K | 编辑 | 2.11 输入 + 14.77 / 输出图片 |
-| grok-imagine-image-pro | — | 编辑 | 2.11 输入 + 14.77 / 输出图片 |
+| grok-imagine-image-2.0 | 1K | 编辑 | 2.11–6.33 输入 + 12.66(中等)/ 8.44(低)每张输出图片 |
+| | 2K | 编辑 | 2.11–6.33 输入 + 16.88(中等)/ 12.66(低)每张输出图片 |
+| grok-imagine-image-quality | 1K | 编辑 | 2.11–6.33 输入 + 8.44 / 输出图片 |
+| | 2K | 编辑 | 2.11–6.33 输入 + 12.66 / 输出图片 |
+| grok-imagine-image-pro | 1K | 编辑 | 2.11 输入 + 8.44 / 输出图片 |
+| | 2K | 编辑 | 2.11 输入 + 12.66 / 输出图片 |
| grok-imagine-image | — | 编辑 | 0.422–1.266 输入 + 4.22 / 输出图片 |
+非 pro 的编辑模型按每张输入图 **2.11 积分** 计费,输入图为 **1 到 3** 张(**2.11–6.33**)。`grok-imagine-image` 为每张输入图 **0.422 积分**(**0.422–1.266**)。pro 编辑行只按一张输入图 **2.11 积分** 计费。
+
+`grok-imagine-image-quality` 与 `grok-imagine-image-pro` 已被 xAI 停用。图像节点仍会列出这两个模型,并按 `grok-imagine-image-2.0` 的低档费率计费。
+
### 视频
| 模型 | 分辨率 | 积分 |
@@ -1772,6 +1961,8 @@ Pruna P-Video-2 按生成视频的秒数计费。总积分 = **(积分 / 秒) ×
| | 720p | 21.12 / 秒 输出 + 0.603 / 参考图片 |
| grok-imagine-video (扩展)| — | 6.03–45.26 输入 + 15.09 × 延长秒 输出 / 次 |
+**扩展** 按每秒延长 **15.09 积分** 计费,并另加源视频的输入费用:2 秒源视频为 **6.03 积分**,15 秒源视频为 **45.26 积分**(源视频必须是 2–15 秒)。线上徽章看不到源视频时长,因此显示 **6.03–45.26** 的输入区间,再加上按秒计算的延长费用。
+
## 云端 GPU
| GPU 类型 | 积分 |