feat: OpenAI/compatible models (such as Azure OpenAI) can declare if they use the response API instead via use_openai_responses_api

providers/openai_compatible.py

@@ -597,8 +597,16 @@ class OpenAICompatibleProvider(ModelProvider):
                 completion_params[key] = value
 
         # Check if this model needs the Responses API endpoint
-        # Both o3-pro and gpt-5-codex use the new Responses API
-        if resolved_model in ["o3-pro", "gpt-5-codex"]:
+        # Prefer capability metadata; fall back to static map when capabilities unavailable
+        use_responses_api = False
+        if capabilities is not None:
+            use_responses_api = getattr(capabilities, "use_openai_response_api", False)
+        else:
+            static_capabilities = self.get_all_model_capabilities().get(resolved_model)
+            if static_capabilities is not None:
+                use_responses_api = getattr(static_capabilities, "use_openai_response_api", False)
+
+        if use_responses_api:
             # These models require the /v1/responses endpoint for stateful context
             # If it fails, we should not fall back to chat/completions
             return self._generate_with_responses_endpoint(

providers/openai_provider.py

@@ -135,6 +135,7 @@ class OpenAIModelProvider(OpenAICompatibleProvider):
             temperature_constraint=TemperatureConstraint.create("fixed"),
             description="Professional-grade reasoning (200K context) - EXTREMELY EXPENSIVE: Only for the most complex problems requiring universe-scale complexity analysis OR when the user explicitly asks for this model. Use sparingly for critical architectural decisions or exceptionally complex debugging that other models cannot handle.",
             aliases=["o3pro"],
+            use_openai_response_api=True,
         ),
         "o4-mini": ModelCapabilities(
             provider=ProviderType.OPENAI,
@@ -191,6 +192,7 @@ class OpenAIModelProvider(OpenAICompatibleProvider):
             temperature_constraint=TemperatureConstraint.create("range"),
             description="GPT-5 Codex (400K context) Specialized for coding, refactoring, and software architecture.",
             aliases=["gpt5-codex", "codex", "gpt-5-code", "gpt5-code"],
+            use_openai_response_api=True,
         ),
     }
 

providers/shared/model_capabilities.py

@@ -50,6 +50,7 @@ class ModelCapabilities:
     supports_images: bool = False
     supports_json_mode: bool = False
     supports_temperature: bool = True
+    use_openai_response_api: bool = False
 
     # Additional attributes
     max_image_size_mb: float = 0.0