fix(zen): read nested usage in non-stream responses

Extract usage from each provider's actual response shape so OpenAI Responses payloads nested under response.usage do not crash Zen's non-stream tracking path.

packages/console/app/test/provider-usage.test.ts

@@ -0,0 +1,79 @@
+import { describe, expect, test } from "bun:test"
+import { anthropicHelper } from "../src/routes/zen/util/provider/anthropic"
+import { googleHelper } from "../src/routes/zen/util/provider/google"
+import { openaiHelper } from "../src/routes/zen/util/provider/openai"
+import { oaCompatHelper } from "../src/routes/zen/util/provider/openai-compatible"
+
+describe("provider usage extraction", () => {
+  test("reads OpenAI Responses usage from response.usage", () => {
+    const helper = openaiHelper({ reqModel: "gpt-5.4", providerModel: "gpt-5.4" })
+
+    expect(
+      helper.extractBodyUsage({
+        response: {
+          usage: {
+            input_tokens: 13,
+            input_tokens_details: { cached_tokens: 3 },
+            output_tokens: 5,
+            output_tokens_details: { reasoning_tokens: 1 },
+          },
+        },
+      }),
+    ).toEqual({
+      input_tokens: 13,
+      input_tokens_details: { cached_tokens: 3 },
+      output_tokens: 5,
+      output_tokens_details: { reasoning_tokens: 1 },
+    })
+  })
+
+  test("reads Anthropic usage from message.usage", () => {
+    const helper = anthropicHelper({ reqModel: "claude-sonnet", providerModel: "claude-sonnet-4-5" })
+
+    expect(
+      helper.extractBodyUsage({
+        message: {
+          usage: {
+            input_tokens: 10,
+            output_tokens: 4,
+          },
+        },
+      }),
+    ).toEqual({
+      input_tokens: 10,
+      output_tokens: 4,
+    })
+  })
+
+  test("reads OA-compatible usage from usage", () => {
+    const helper = oaCompatHelper({ reqModel: "gpt-4o-mini", providerModel: "gpt-4o-mini" })
+
+    expect(
+      helper.extractBodyUsage({
+        usage: {
+          prompt_tokens: 8,
+          completion_tokens: 2,
+        },
+      }),
+    ).toEqual({
+      prompt_tokens: 8,
+      completion_tokens: 2,
+    })
+  })
+
+  test("reads Google usage from usageMetadata", () => {
+    const helper = googleHelper({ reqModel: "gemini-2.5-flash", providerModel: "gemini-2.5-flash" })
+
+    expect(
+      helper.extractBodyUsage({
+        usageMetadata: {
+          promptTokenCount: 11,
+          candidatesTokenCount: 3,
+        },
+      }),
+    ).toEqual({
+      promptTokenCount: 11,
+      candidatesTokenCount: 3,
+    })
+  })
+})