diff --git a/docs/content/advanced/model-configuration.md b/docs/content/advanced/model-configuration.md
index df10b628ae..216e7e1de9 100644
--- a/docs/content/advanced/model-configuration.md
+++ b/docs/content/advanced/model-configuration.md
@@ -689,6 +689,18 @@ This is the load-time reasoning configuration. The orthogonal per-request `enabl
 - `reasoning_effort: "minimal" | "low" | "medium" | "high" | "xhigh" | "max"` enables thinking, unless the model config explicitly set `reasoning.disable: true` (an operator's explicit disable wins and is never re-enabled by a request).
 {{% /notice %}}
 
+The `/v1/responses` endpoint accepts the same effort levels through `reasoning.effort`.
+It also forwards request `metadata` to the backend, including `enable_thinking` overrides.
+
+```json
+{
+  "model": "qwen3",
+  "input": "Say hello",
+  "reasoning": { "effort": "none" },
+  "metadata": { "enable_thinking": "false" }
+}
+```
+
 #### `reasoning_effort` as a chat-template kwarg
 
 `reasoning_effort` is also forwarded to the backend as a `chat_template_kwarg`, so models whose **jinja chat template** keys on it - e.g. gpt-oss (Harmony) or LFM2.5 - honor the **level**, not just the on/off `enable_thinking` flag. This matters for models that ignore `enable_thinking` entirely (LFM2.5 keeps emitting `<think>` for `enable_thinking=false`, but respects `reasoning_effort`).