From 7f5275ed6bf8e3278a6a4e5b3f2ae7eec63c9592 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rouzbeh=E2=80=A0?=
 <78313022+rqzbeh@users.noreply.github.com>
Date: Mon, 17 Aug 2026 14:58:03 +0330
Subject: [PATCH] fix(usage): surface Gemini cachedContentTokenCount as
 cached_tokens (#10465)

* fix(usage): read Gemini usageMetadata out of the antigravity response envelope

Port decolua/9router#59d858b: antigravity/gemini-cli wrap non-streaming
payloads in { response: {...} }, so extractUsageFromResponse only saw the
top-level usageMetadata and every non-streaming antigravity request logged
zero usage (IN 0 | OUT 0) and zeroed usage-dashboard rows. Top-level
metadata keeps priority; OpenAI/Claude branches untouched.

* chore(changelog): fragment for #10430 antigravity usage envelope

* fix(usage): surface Gemini cachedContentTokenCount as cached_tokens

Review follow-up on #10430: the Gemini branch of extractUsageFromResponse
ignored cachedContentTokenCount, so non-streaming cache-hit tokens never
reached the cached_tokens field the OpenAI/Claude/Responses branches
already populate (and the streaming path surfaces at usageTracking.ts:684).

Adds cached_tokens: usageMetadata.cachedContentTokenCount || 0, updates
the three Gemini assertions (envelope fixture already carried
cachedContentTokenCount: 7), and adds a dedicated regression test.

* chore(changelog): fragment for #10465 Gemini cached_tokens surfacing

* test(fix): refresh expired alibaba quota sample validity and onnxruntime pin for v3.8.50 base

- alibaba-free-tier-quota-fetcher.test.ts: sample quotaValidityPeriod
  (2026-08-16 16:00 UTC) is in the past, making every quota entry classify
  as expired/not_capable; bump to 2028-01-01 UTC so the text/merge
  classification tests exercise the intended path again.
- optional-transformers-dependency.test.ts: onnxruntime-node pin assertion
  updated from ~1.24.3 to ~1.27.0 to match package.json (bumped by #10403);
  the regular-not-optional intent is unchanged.

---------

Co-authored-by: Rouzbeh <rqzbeh@users.noreply.github.com>
---
 .../fixes/10465-gemini-cached-tokens.md       |  1 +
 open-sse/handlers/usageExtractor.ts           |  1 +
 tests/unit/usage-extractor.test.ts            | 26 +++++++++++++++++++
 3 files changed, 28 insertions(+)
 create mode 100644 changelog.d/fixes/10465-gemini-cached-tokens.md

diff --git a/changelog.d/fixes/10465-gemini-cached-tokens.md b/changelog.d/fixes/10465-gemini-cached-tokens.md
new file mode 100644
index 0000000000..0acd31720a
--- /dev/null
+++ b/changelog.d/fixes/10465-gemini-cached-tokens.md
@@ -0,0 +1 @@
+- **fix(usage):** surface Gemini `cachedContentTokenCount` into `cached_tokens` for non-streaming requests so cache-hit accounting matches the OpenAI/Claude/Responses branches and the streaming path (follow-up to the #10430 envelope fix) ([#10465](https://github.com/diegosouzapw/OmniRoute/pull/10465)) — thanks @rqzbeh
diff --git a/open-sse/handlers/usageExtractor.ts b/open-sse/handlers/usageExtractor.ts
index 313d1ff0e4..f424996ca4 100644
--- a/open-sse/handlers/usageExtractor.ts
+++ b/open-sse/handlers/usageExtractor.ts
@@ -104,6 +104,7 @@ export function extractUsageFromResponse(responseBody, provider) {
     return {
       prompt_tokens: usageMetadata.promptTokenCount || 0,
       completion_tokens: (usageMetadata.candidatesTokenCount || 0) + thoughts,
+      cached_tokens: usageMetadata.cachedContentTokenCount || 0,
       reasoning_tokens: thoughts,
     };
   }
diff --git a/tests/unit/usage-extractor.test.ts b/tests/unit/usage-extractor.test.ts
index 851032690c..af328e52b6 100644
--- a/tests/unit/usage-extractor.test.ts
+++ b/tests/unit/usage-extractor.test.ts
@@ -228,6 +228,7 @@ test("extractUsageFromResponse reads Gemini usageMetadata and thinking tokens",
   assert.deepEqual(usage, {
     prompt_tokens: 11,
     completion_tokens: 7,
+    cached_tokens: 0,
     reasoning_tokens: 2,
   });
 });
@@ -252,6 +253,7 @@ test("extractUsageFromResponse reads Gemini usageMetadata from the antigravity r
   assert.deepEqual(usage, {
     prompt_tokens: 42,
     completion_tokens: 17,
+    cached_tokens: 7,
     reasoning_tokens: 4,
   });
 });
@@ -268,10 +270,34 @@ test("extractUsageFromResponse prefers top-level usageMetadata over the envelope
   assert.deepEqual(usage, {
     prompt_tokens: 1,
     completion_tokens: 2,
+    cached_tokens: 0,
     reasoning_tokens: 0,
   });
 });
 
+test("extractUsageFromResponse surfaces Gemini cachedContentTokenCount as cached_tokens", () => {
+  // Review follow-up on #10430: match the OpenAI/Claude/Responses branches and
+  // the streaming path (usageTracking.ts) by surfacing Gemini cache-hit tokens.
+  const usage = extractUsageFromResponse(
+    {
+      usageMetadata: {
+        promptTokenCount: 30,
+        candidatesTokenCount: 10,
+        thoughtsTokenCount: 3,
+        cachedContentTokenCount: 12,
+      },
+    },
+    "gemini"
+  );
+
+  assert.deepEqual(usage, {
+    prompt_tokens: 30,
+    completion_tokens: 13,
+    cached_tokens: 12,
+    reasoning_tokens: 3,
+  });
+});
+
 test("extractUsageFromResponse returns null when usage is missing", () => {
   const usage = extractUsageFromResponse(
     {