diff --git a/src/adapters/anthropic.ts b/src/adapters/anthropic.ts
index 7d49e7ca70..19a6747d5a 100644
--- a/src/adapters/anthropic.ts
+++ b/src/adapters/anthropic.ts
@@ -935,9 +935,10 @@ export function createAnthropicAdapter(provider: OcxProviderConfig, cacheRetenti
           // output, so high effort needs the same total-token headroom as budget thinking or a
           // default 8192-token request can spend everything on thought and return empty text.
           body.thinking = { type: "adaptive" };
-          body.output_config = { effort: adaptiveEffort(parsed.options.reasoning) };
+          const effort = adaptiveEffort(parsed.options.reasoning);
+          body.output_config = { effort };
           const explicitMaxOut = parsed.options.maxOutputTokens;
-          const wantBudget = reasoningBudget(parsed.options.reasoning);
+          const wantBudget = reasoningBudget(effort);
           const floor = wantBudget + OUTPUT_HEADROOM;
           // Preserve explicit caller limits as-is; for omitted limits use the adaptive ceiling
           // so effort=max (budget=32k) still leaves OUTPUT_HEADROOM tokens for visible output.
diff --git a/tests/anthropic-reasoning.test.ts b/tests/anthropic-reasoning.test.ts
index ab91ee8f05..7e31501a6c 100644
--- a/tests/anthropic-reasoning.test.ts
+++ b/tests/anthropic-reasoning.test.ts
@@ -68,6 +68,7 @@ describe("anthropic extended-thinking gate", () => {
   test("adaptive-thinking model maps unsupported 'minimal' effort to 'low'", async () => {
     const b = await bodyOf(parsed("minimal", {}, "claude-fable-5"));
     expect(b.output_config).toEqual({ effort: "low" });
+    expect(b.max_tokens).toBe(12_288);
   });
 
   test("forwards Responses JSON Schema output format to Anthropic", async () => {