Model returned an empty response
Sometimes, chat-models may work correctly and give a successful response, but that response is empty. In most cases, this is caused by the max_tokens
(Output length) parameter being too low. It happens because some models "think" before giving a response, and they may hit this limit while thinking, thus not having any room left for a response.
To resolve the issue, try this:
- If it's a Playground request, increase the Output length value on the right side of the screen
- If it's an API request, increase the
max_tokens
value. If you are not setting it in your request, then add it, and use a number higher than default (512), like so:
"max_tokens": 3000
If this did not help, please contact our technical support.