The new o1 series of models deprecate the max_tokens parameter in favor of a new max_completion_tokens parameter and I’d like to understand the rational for this change as it’s likely to have wide sweeping impacts for what appears to be a simple wording change. Most of the breaking changes up to th…

Hi, Atty from OpenAI here — max_tokens continues to be supported in all existing models, but the o1 series only supports max_completion_tokens. We are doing this because max_tokens previously meant both the number of tokens we generated (and billed you for) and the number of tokens you got back in …

Why was max_tokens changed to max_completion_tokens?

API Feedback

stevenic September 13, 2024, 6:27pm 6

If anything you should have just added a new max_reasoning_tokens parameter to let the user constrain how long o1 is allowed to think.

Topic		Replies	Views
Introducing OpenAI o1-preview \| New OpenAI Announcement API announcement , chatgpt , news	39	5758	September 17, 2024
Confused about max_tokens - parameter with GTP4-turbo (128k-tokenUsedForPrompt or 4K) API	16	6897	May 11, 2024
Max_tokens seems to do nothing for me 3.5 Turbo API	14	3306	December 18, 2023
Not allowed to have all 8192 tokens API gpt-4	16	11178	December 18, 2023
Get all requested max tokens with gpt-3.5-turbo-instruct API gpt-35-turbo-instruc	20	7374	January 21, 2024

Why was max_tokens changed to max_completion_tokens?

Related topics