GPT-5 not supported max_tokens must be max_completion_tokens

[netcorefactory]

Bug description
When using gpt-5 i get the following error:
Finished call to 'metagpt.actions.action_node.ActionNode._aask_v1' after 16.915(s), this was the 6th time calling it. exp: 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'}}

Bug solved method
Used latest docker image

Environment information
windows 11

• LLM type and model name: GPT-5
• System version: docker image
• Python version:
• MetaGPT version or branch: main

Screenshots or logs
Traceback (most recent call last):
File "/app/metagpt/metagpt/utils/common.py", line 650, in wrapper
result = await func(self, *args, **kwargs)
File "/app/metagpt/metagpt/team.py", line 134, in run
await self.env.run()
openai.BadRequestError: 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'}}

[dickymoore]

I ran into the same problem when trying to use GPT-5 (max_tokens not supported). I’ve opened PR #1883
which updates the OpenAI provider to use max_completion_tokens instead.

Confirmed working with GPT-5 (gpt-5 model)

Backward compatibility tested with GPT-4 (gpt-4-turbo)

I’d really appreciate it if others could test this branch in their environments and report back any issues. Feedback on the implementation is also very welcome.

Thanks!

[github-actions]

This issue has no activity in the past 30 days. Please comment on the issue if you have anything to add.