BadRequestError: 'max_tokens' parameter deprecated - should use 'max_completion_tokens'

[ryota-murakami]

Bug Report

Environment

• Package version: 0.10.0
• Node version: 22.18.0 (via Volta)
• Error location: openai/core/error.mjs:41

Error Message

file:///Users/ryotamurakami/.volta/tools/image/packages/@laststance/git-gpt-commit/lib/node_modules/@laststance/git-gpt-commit/node_modules/openai/core/error.mjs:41
            return new BadRequestError(status, error, message, headers);
                   ^

BadRequestError: 400 Unsupported parameter: 'max_tokens' is not supported with this model. Use 'max_completion_tokens' instead.

Description

OpenAI has deprecated the max_tokens parameter in newer API versions and models (particularly the o1 series). The parameter has been renamed to max_completion_tokens.

Suggested Fix

Replace all instances of max_tokens with max_completion_tokens in the OpenAI API calls.

Temporary Workaround

Users experiencing this issue can:

1. Use an older model that still supports max_tokens
2. Downgrade to an older version of the package
3. Wait for the fix to be released

References

• OpenAI Developer Forum Discussion