Error with model selection with openai reasoning models

[Pythonista7]

I notice that OpenAIAugmentedLLM on init sets the self._reasoning flag , the default fallback model is gpt-4o but when i specifically pass a model_name which is a reasoning model during generate_str(request_params=...) it fails with the below error.

[mcp_agent.workflows.llm.augmented_llm_openai.writing_assistant] Error: Error code: 400 - {'error': {'message': "Unsupported parameter: 
'max_tokens' is not supported with this model. Use 'max_completion_tokens' instead.", 'type': 'invalid_request_error', 'param': 'max_tokens',
 'code': 'unsupported_parameter'}}

I believe we would need to update the self._reasoning after the self.select_model() returns (line: 154) here?

If there is an alternate way to resolve this or if a code change would be required would be happy to implement them based on suggestion :) Thanks!

[StreetLamb]

Hi @Pythonista7, thanks for raising the issue! If you would like to, another thing to resolve would be changing max_tokens to max_completion_tokens since max_tokens has been deprecated.

[Pythonista7]

@StreetLamb great suggestion! Have update the PR with the change: #139