diff --git a/CHANGELOG.md b/CHANGELOG.md
index b9fb460..ccea000 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -2,6 +2,10 @@
 
 All notable changes are recorded here. Versions follow [SemVer](https://semver.org/).
 
+## v0.7.11 (2026-08-23)
+
+- **P0 cache-billing fix (passthrough path)** — PR #128 (v0.7.10) only patched the cross-protocol SSE conversion path; the same-protocol passthrough path (`UsageCapture::finish()` in `src/gateway.rs`, used when client and upstream speak the same protocol, e.g. openai→openai) still passed the **full** `prompt_tokens` as input (including cache-hit tokens) and only recognized OpenAI's `prompt_tokens_details.cached_tokens` spelling — so DeepSeek responses double-counted cached tokens into the total (~2× tokens, per-call uncached showed 241,776 instead of 112). `finish()` now extracts cached via all three spellings (DeepSeek native `prompt_cache_hit_tokens` → OpenAI `prompt_tokens_details.cached_tokens` → Anthropic `cache_read_input_tokens`) and returns `(input − cached).max(0.0)` disjoint, matching the `sse.rs` logic; tests extended with cached cases for all three spellings (rant 2026-08-23T14:05:02, PR #130)
+
 ## v0.7.10 (2026-08-23)
 
 - **Request body limit raised to 70MB** — axum's default 2MB body limit rejected long LLM contexts (~1M token) and large image-base64 payloads with 413; gateway now applies `RequestBodyLimitLayer::new(70 * 1024 * 1024)` (rant 2026-08-22T23:20:00)
diff --git a/Cargo.lock b/Cargo.lock
index 4e98692..7b6c3db 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -51,7 +51,7 @@ dependencies = [
 
 [[package]]
 name = "aitokenpool"
-version = "0.7.10"
+version = "0.7.11"
 dependencies = [
  "aes-gcm",
  "anyhow",
diff --git a/Cargo.toml b/Cargo.toml
index 1bc654e..a79cd7a 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "aitokenpool"
-version = "0.7.10"
+version = "0.7.11"
 edition = "2021"
 description = "AI Token 共享池 — 企业 key 池 + 公共共享市场"
 license = "MIT"
diff --git a/ui/index.html b/ui/index.html
index 47244fc..04c829c 100644
--- a/ui/index.html
+++ b/ui/index.html
@@ -5,7 +5,7 @@
 <meta name="viewport" content="width=device-width, initial-scale=1.0">
 <title>AITokenPool</title>
 <link rel="icon" type="image/svg+xml" href="data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 64 64'%3E%3Cdefs%3E%3ClinearGradient id='g' x1='0' y1='0' x2='1' y2='1'%3E%3Cstop offset='0' stop-color='%234ecdc4'/%3E%3Cstop offset='1' stop-color='%232a9d8f'/%3E%3C/linearGradient%3E%3C/defs%3E%3Crect width='64' height='64' rx='14' fill='url(%23g)'/%3E%3Ctext x='32' y='43' font-family='-apple-system,Segoe UI,Arial,sans-serif' font-size='30' font-weight='800' text-anchor='middle' fill='%23062021'%3EAT%3C/text%3E%3C/svg%3E">
-<link rel="stylesheet" href="css/style.css?v=20260823-1">
+<link rel="stylesheet" href="css/style.css?v=20260823-2">
 </head>
 <body>
 
@@ -682,9 +682,9 @@ <h3 id="chat-title">使用模型</h3>
   </div>
 </div>
 
-<script src="js/api.js?v=20260823-1"></script>
-<script src="js/data.js?v=20260823-1"></script>
-<script src="js/i18n.js?v=20260823-1"></script>
-<script src="js/app.js?v=20260823-1"></script>
+<script src="js/api.js?v=20260823-2"></script>
+<script src="js/data.js?v=20260823-2"></script>
+<script src="js/i18n.js?v=20260823-2"></script>
+<script src="js/app.js?v=20260823-2"></script>
 </body>
 </html>