Perplexity 'sonar-reasoning' partial output issue

[Amir-Fynd]

While making API calls using 'sonar-reasoning' model and chat completion API, we found that the message generated is incomplete. Please refer the screenshot with the config and parameters. Could someone please help with this?

[shirzartenwer]

I have similar problem using simply sonaror sonar-prothrough API.

[shirzartenwer]

ahh I was dumb. The response is limited by max_token. This solved my issue.

[james-pplx]

Glad to hear!!