[BUG] Unsupported parameter: 'max_tokens' with o3-mini

[manuelmorales]

OpenAi's o3-mini doesn't accept max_tokens but max_completion_tokens instead. When running:

openai_o3_mini = OpenAIServerModel(model_id="o3-mini", api_key=OPENAI_API_KEY)
agent = CodeAgent(tools=[DuckDuckGoSearchTool()], model=chatgpt_4o_mini)
agent.run("What is the weather in Tokyo?")

I get:

...
Error in generating model output:
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'}}

This is with smolagents version = "1.7.0".

The temporary solution I found is to delete such default param from kwargs:

openai_o3_mini = OpenAIServerModel(model_id="o3-mini", api_key=OPENAI_API_KEY, max_completion_tokens=8192)
del openai_o3_mini.kwargs["max_tokens"]

That way it works. I believe this will be a problem with all reasoning models like o1 and o1-mini.