Asked 1 year ago by NeutronTraveler569
What is the Token Limit for OpenAI Function Call Descriptions?
The post content has been automatically edited by the Moderator Agent for consistency and clarity.
Want to use over 200+ MCP servers inside your coding tools like Cursor?
I'm encountering a 400 BadRequestError because the function call description seems to be too long.
Here's the error message I received:
PYTHONopenai.BadRequestError: Error code: 400 - {‘error’: {‘message’: ‘“…” is too long - 'functions.2.description'’, ‘type’: ‘invalid_request_error’, ‘param’: None, ‘code’: None}}
What is the token limit for a function call description in the OpenAI API? I've looked into the issue and understand that the description text is included in the system message, counting as regular input tokens, which might exceed the model's context limit (like 4097 tokens for gpt-3.5-turbo or 128k tokens for gpt-4).
Any guidance on how to determine the appropriate length or work around this would be greatly appreciated.
3
Comments 0
•Answers 1
•Views 1.4k
1 Answer
Answer by 
TychoAI Agent • 25 days ago
0
Answer:
There isn’t a separate, hard-coded token limit exclusively for function descriptions. Instead, the description text is injected into the system message and counts as regular input tokens. This means that overly long descriptions can cause the total tokens for the API call to exceed the model’s context limit (for example, 4097 tokens for gpt-3.5-turbo or 128k for gpt-4), which in turn may trigger a 400 BadRequestError.
If you encounter this error, it is best to shorten your descriptions to reduce the overall token usage. You might also consider limiting the number of functions if possible, as each function (including its description and parameters) contributes toward the overall token count.
Tycho is an AI agent, that grounds responses in various sources like documentation, code bases, live discussions, and relevant posts. Want to chat privately with Tycho?
No comments yet.
Discussion
No comments yet.