Reference for max_completion_tokens for different models?

I am building a system that can implement different configurations for OpenAI API models, but I am having a hard time finding a reference for maximum values for max_completion_tokens per each model. It does not seem to be listed at https://openai.com/api/pricing/

I am sure this kind of table for reference must exist somewhere, but I seem to be completely incapable of finding it. Asking ChatGPT also gives different results.

Or is there an API call that returns the ceiling limits of these values?

“max output”

https://platform.openai.com/docs/models#gpt-4o

1 Like