fix(github-copilot): route per-model on /v1/responses based on model info

litellm/llms/github_copilot/responses/transformation.py

@@ -22,6 +22,7 @@
 )
 from litellm.types.router import GenericLiteLLMParams
 from litellm.types.utils import LlmProviders
+from litellm.utils import _get_model_info_helper
 
 from ..authenticator import Authenticator
 from ..common_utils import (
@@ -38,6 +39,53 @@
     LiteLLMLoggingObj = Any
 
 
+def github_copilot_supports_responses_api(model: str) -> bool:
+    """
+    Resolve whether to use the native Responses API for a github_copilot model.
+
+    Copilot's /v1/responses endpoint is per-model (only some models like
+    gpt-5.5, gpt-5.4, gpt-5.4-mini opt in upstream). The Responses API config
+    is registered for github_copilot provider-wide, so this function gates
+    per-model to keep /v1/responses calls from failing upstream on chat-only
+    Copilot models like claude-opus-4.7 or gemini-3.1-pro-preview.
+
+    The router calls ``litellm.register_model`` for every proxy deployment,
+    which merges the user's per-deployment ``model_info`` (e.g. ``mode: chat``
+    to force a chat-only override) into ``litellm.model_cost`` before any
+    request runs. ``_get_model_info_helper`` therefore returns merged data
+    with user overrides already applied.
+
+    Resolution order (first match wins):
+      1. ``mode == "responses"`` → True (positive opt-in; user or catalog).
+      2. ``mode == "chat"`` → False (explicit opt-out wins over endpoint
+         declarations, letting users force the bridge for dual-endpoint models).
+      3. ``"/v1/responses"`` in ``supported_endpoints`` → True.
+      4. Otherwise → False (conservative default; the bridge always works
+         because every Copilot model supports /chat/completions).
+
+    Catalog lookup raising (model not registered) → False (conservative).
+    """
+    try:
+        info = _get_model_info_helper(model=model, custom_llm_provider="github_copilot")
+    except Exception as e:
+        verbose_logger.debug(
+            "github_copilot_supports_responses_api: get_model_info failed "
+            "for %s: %s",
+            model,
+            e,
+        )
+        return False
+
+    mode = info.get("mode")
+    if mode == "responses":
+        return True
+    if mode == "chat":
+        return False
+
+    endpoints = info.get("supported_endpoints")
+    return isinstance(endpoints, list) and "/v1/responses" in endpoints
+
+
 class GithubCopilotResponsesAPIConfig(OpenAIResponsesAPIConfig):
     """
     Configuration for GitHub Copilot's Responses API.

litellm/utils.py

@@ -8575,7 +8575,13 @@ def _get_python_responses_api_config(
         elif litellm.LlmProviders.XAI == provider:
             return litellm.XAIResponsesAPIConfig()
         elif litellm.LlmProviders.GITHUB_COPILOT == provider:
-            return litellm.GithubCopilotResponsesAPIConfig()
+            from litellm.llms.github_copilot.responses.transformation import (
+                github_copilot_supports_responses_api,
+            )
+
+            if model is None or github_copilot_supports_responses_api(model=model):
+                return litellm.GithubCopilotResponsesAPIConfig()
+            return None
         elif litellm.LlmProviders.CHATGPT == provider:
             return litellm.ChatGPTResponsesAPIConfig()
         elif litellm.LlmProviders.LITELLM_PROXY == provider: