gitlab-ai-provider 6.7.0 → 6.7.1

package/CHANGELOG.md

@@ -2,6 +2,13 @@
 
 All notable changes to this project will be documented in this file. See [Conventional Commits](https://conventionalcommits.org) for commit guidelines.
 
+## <small>6.7.1 (2026-05-26)</small>
+
+- Merge branch 'fix/anthropic-cache-token-counting' into 'main' ([298e2d0](https://gitlab.com/vglafirov/gitlab-ai-provider/commit/298e2d0))
+- build: regenerate dist for anthropic cache token fix ([627d6f7](https://gitlab.com/vglafirov/gitlab-ai-provider/commit/627d6f7))
+- test(anthropic): add streaming test for cache token handling ([6033350](https://gitlab.com/vglafirov/gitlab-ai-provider/commit/6033350))
+- fix(anthropic): include cache tokens in input token total ([699af0a](https://gitlab.com/vglafirov/gitlab-ai-provider/commit/699af0a))
+
 ## 6.7.0 (2026-05-18)
 
 - Merge branch 'feat/add-gpt-5-5-model-mapping' into 'main' ([10e84d1](https://gitlab.com/vglafirov/gitlab-ai-provider/commit/10e84d1))

package/dist/index.js

@@ -474,14 +474,15 @@ ${message.content}` : message.content;
     return { unified, raw: stopReason ?? void 0 };
   }
   createUsage(params) {
-    const inputTotal = params?.inputTotal;
+    const inputNoCache = params?.inputTotal;
     const outputTotal = params?.outputTotal;
     const cacheRead = params?.cacheRead;
     const cacheWrite = params?.cacheWrite;
+    const inputTotal = inputNoCache != null ? inputNoCache + (cacheRead ?? 0) + (cacheWrite ?? 0) : void 0;
     return {
       inputTokens: {
         total: inputTotal,
-        noCache: inputTotal == null ? void 0 : Math.max(0, inputTotal - (cacheRead ?? 0) - (cacheWrite ?? 0)),
+        noCache: inputNoCache,
         cacheRead,
         cacheWrite
       },
@@ -609,12 +610,13 @@ ${message.content}` : message.content;
                 switch (event.type) {
                   case "message_start":
                     if (event.message.usage) {
+                      const msgUsage = event.message.usage;
                       usage = self.createUsage({
-                        inputTotal: event.message.usage.input_tokens,
+                        inputTotal: msgUsage.input_tokens,
                         outputTotal: usage.outputTokens.total,
                         outputReasoning: usage.outputTokens.reasoning,
-                        cacheRead: usage.inputTokens.cacheRead,
-                        cacheWrite: usage.inputTokens.cacheWrite,
+                        cacheRead: msgUsage.cache_read_input_tokens,
+                        cacheWrite: msgUsage.cache_creation_input_tokens,
                         raw: usage.raw
                       });
                     }
@@ -688,12 +690,13 @@ ${message.content}` : message.content;
                   }
                   case "message_delta":
                     if (event.usage) {
+                      const deltaUsage = event.usage;
                       usage = self.createUsage({
-                        inputTotal: usage.inputTokens.total,
-                        outputTotal: event.usage.output_tokens,
+                        inputTotal: usage.inputTokens.noCache,
+                        outputTotal: deltaUsage.output_tokens,
                         outputReasoning: usage.outputTokens.reasoning,
-                        cacheRead: usage.inputTokens.cacheRead,
-                        cacheWrite: usage.inputTokens.cacheWrite,
+                        cacheRead: deltaUsage.cache_read_input_tokens ?? usage.inputTokens.cacheRead,
+                        cacheWrite: deltaUsage.cache_creation_input_tokens ?? usage.inputTokens.cacheWrite,
                         raw: usage.raw
                       });
                     }
@@ -1681,7 +1684,7 @@ var GitLabOpenAILanguageModel = class {
 var import_isomorphic_ws = __toESM(require("isomorphic-ws"));
 
 // src/version.ts
-var VERSION = true ? "6.6.0" : "0.0.0-dev";
+var VERSION = true ? "6.7.0" : "0.0.0-dev";
 
 // src/gitlab-workflow-types.ts
 var WorkflowType = /* @__PURE__ */ ((WorkflowType2) => {