OpenRouter stop cap is 4, GGUF tool-loop final pass forwards new fields

OpenRouter normalises to OpenAI's chat schema and inherits the 4-entry stop cap. The default 16-cap was too permissive; add stop_max=4 on both the backend provider registry and the frontend PROVIDER_STOP_MAX map.

The GGUF tool-iteration final-answer pass at llama_cpp.py:5182 was carrying only the legacy sampling fields. Forward frequency_penalty, seed, and parallel_tool_calls there too so the cap-exhausted path matches the per-iteration loop.

Test pins the OpenRouter 4-cap.

studio/backend/core/inference/llama_cpp.py

@@ -5202,6 +5202,14 @@ class LlamaCppBackend:
         stream_payload["t_max_predict_ms"] = _DEFAULT_T_MAX_PREDICT_MS
         if stop:
             stream_payload["stop"] = stop
+        # Match the per-iteration tool loop above so sampling behavior
+        # stays consistent when the cap-exhausted final-answer pass runs.
+        if frequency_penalty is not None:
+            stream_payload["frequency_penalty"] = frequency_penalty
+        if seed is not None:
+            stream_payload["seed"] = seed
+        if parallel_tool_calls is not None:
+            stream_payload["parallel_tool_calls"] = parallel_tool_calls
         stream_payload["stream_options"] = {"include_usage": True}
 
         cumulative = ""

studio/backend/core/inference/providers.py

@@ -312,6 +312,9 @@ PROVIDER_REGISTRY: dict[str, dict[str, Any]] = {
         },
         "notes": "Unified gateway to 300+ models across all major providers. HTTP-Referer and X-Title headers sent for attribution.",
         "model_list_mode": "curated",
+        # OpenRouter normalises to OpenAI's chat schema and inherits
+        # the 4-entry stop cap.
+        "stop_max": 4,
     },
 }
 

studio/backend/tests/test_sampling_params_routing.py

@@ -607,6 +607,34 @@ def test_kimi_web_search_uses_kimi_stop_cap_5(monkeypatch):
     assert body["stop"] == ["S0", "S1", "S2", "S3", "S4"], body
 
 
+def test_openrouter_stop_cap_is_4(monkeypatch):
+    """OpenRouter normalises to OpenAI's chat schema and inherits the
+    4-entry stop cap; the default 16-cap is too permissive for it."""
+    captured = _install_mock(monkeypatch, sse_payload = _oai_done_payload())
+
+    async def run():
+        client = ExternalProviderClient(
+            provider_type = "openrouter",
+            base_url = "https://openrouter.ai/api/v1",
+            api_key = "or-test",
+        )
+        async for _ in client.stream_chat_completion(
+            messages = [{"role": "user", "content": "hi"}],
+            model = "openai/gpt-4o",
+            temperature = 0.5,
+            top_p = 0.9,
+            max_tokens = 64,
+            stop = [f"S{i}" for i in range(10)],
+        ):
+            pass
+        await client.close()
+
+    _drive(run())
+    body = captured["body"]
+    assert len(body.get("stop", [])) == 4, body
+    assert body["stop"] == ["S0", "S1", "S2", "S3"], body
+
+
 def test_kimi_drops_stop_strings_over_32_bytes(monkeypatch):
     """Kimi limits each stop string to <= 32 bytes per
     https://platform.kimi.ai/docs/api/chat. Drop overlong entries

studio/frontend/src/features/chat/provider-capabilities.ts

@@ -64,12 +64,13 @@ export interface ProviderCapabilities {
  * Per-provider stop-sequence max count. Resolved by
  * `getProviderStopMax(providerType)`. Mirrors the backend's
  * `provider_info.stop_max` for the same provider type.
- *   - openai:    4   (Chat Completions hard cap; Responses drops stop)
- *   - anthropic: 16  (client-side guard; docs publish no max)
- *   - kimi:      5   (https://platform.kimi.ai/docs/api/chat)
- *   - deepseek:  16  (https://api-docs.deepseek.com/api/create-chat-completion)
- *   - mistral:   16  (no documented max; widen to permissive default)
- *   - default:   16  (covers ollama, vllm, llama.cpp, openrouter, custom)
+ *   - openai:     4   (Chat Completions hard cap; Responses drops stop)
+ *   - anthropic:  16  (client-side guard; docs publish no max)
+ *   - kimi:       5   (https://platform.kimi.ai/docs/api/chat)
+ *   - deepseek:   16  (https://api-docs.deepseek.com/api/create-chat-completion)
+ *   - mistral:    16  (no documented max; widen to permissive default)
+ *   - openrouter: 4   (normalises to OpenAI's chat schema)
+ *   - default:    16  (covers ollama, vllm, llama.cpp, custom)
  */
 const PROVIDER_STOP_MAX: Record<string, number> = {
   openai: 4,
@@ -77,6 +78,7 @@ const PROVIDER_STOP_MAX: Record<string, number> = {
   kimi: 5,
   deepseek: 16,
   mistral: 16,
+  openrouter: 4,
 };
 
 export function getProviderStopMax(