Fix model validation and empty choices handling in MiniMaxChatNode

- Validate `model` parameter against MiniMaxChatModel enum before
  forwarding to the API request, preventing invalid model strings.
- Raise RuntimeError instead of returning a fallback string when the
  API returns empty choices, surfacing failures consistently.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: octo-patch

comfy_api_nodes/nodes_minimax.py

@@ -541,7 +541,7 @@ async def execute(
             ApiEndpoint(path="/proxy/minimax/chat/completions", method="POST"),
             response_model=MiniMaxChatResponse,
             data=MiniMaxChatRequest(
-                model=model,
+                model=MiniMaxChatModel(model).value,
                 messages=messages,
                 max_tokens=max_tokens,
                 temperature=temperature,
@@ -550,7 +550,7 @@ async def execute(
 
         if response.choices:
             return IO.NodeOutput(response.choices[0].message.content)
-        return IO.NodeOutput("No response generated by MiniMax model.")
+        raise RuntimeError("No response generated by MiniMax model.")
 
 
 class MinimaxExtension(ComfyExtension):