Use raw Router model ids in API preflight

backend/routes/v1_responses.py

@@ -114,7 +114,7 @@ async def _preflight_hf_router_access(model: str, hf_token: str | None) -> None:
 
     normalized_model = cache_key[1]
     payload = {
-        "model": f"openai/{normalized_model}",
+        "model": normalized_model,
         "messages": [{"role": "user", "content": "Reply with OK."}],
         "max_tokens": 1,
     }