After updating: Network timeout at: https://api.perplexity.ai/chat/completions

[Underpippo]

I have updated my code for the new model:

{
  "model": "sonar",
  "messages": [
    {
      "role": "system",
      "content": "Be precise and concise."
    },
    {
      "role": "user",
      "content": "Private Prompt"
    }
  ]
}

This was my old code:

{
  "model": "llama-3.1-sonar-small-128k-online",
  "messages": [
    {
      "role": "system",
      "content": "Be precise and concise."
    },
    {
      "role": "user",
      "content": "Private Prompt."
    }
  ]
}

As you can see, I have only adapted the model name. Everything worked before, but now I only get a network timeout at: https://api.perplexity.ai/chat/completions

[Daksh-Pepsales]

same issue is with me too!!

[yeshivsher]

Same issue for the sonar model with timeout
but for the sonar-pro I get "unknown"

[yeshivsher]

UPDATE:
it works by removing these properties: max_tokens, temperature (tested for the sonar-pro)

[Underpippo]

As you can see in my code, i dont use these properties

[yeshivsher]

did you try sonar-pro?

[Underpippo]

I have just tested it. Sonar-Pro works perfectly, but is also significantly more expensive, which is why I actually want to use the sonar model

[Underpippo]

I have found my solution:
I accessed Sonar via Retool and the network timeout was automatically canceled due to a timeout setting that was too low.