[BUG] Wrong Perplexity model called (llama 3.0 ->3.1)

[JJsilvera1]

hello. When I use perplexity the call is sending to the wrong model type.

[Service Issue] Perplexity: Bad Request (400): "error": { "message": "Invalid model 'llama-3-sonar-large-32k-chat'. Permitted models can be found in the documentation at https://docs.perplexity.ai/docs/model-cards.", "type": "invalid_model", "code": 400 }

should be 3.1 models.

I just added api access today

[enricoros]

Thanks @JJsilvera1, fixed in Big-AGI-2.

[JJsilvera1]

Okay, thank you!

[JJsilvera1]

@enricoros I signed up for the beta but since I don't have access to it doesn't really help 😂

[enricoros]

Thanks for signing up, I'll personally review your answers.

[enricoros]

@JJsilvera1 - I can have a pre-pilot program and admit you to the 2 alpha server.

V2 is incredibly better than V1, multimodal, resilient, faster, shows token usages, pricings, supports new architectures, and a few new large features I'm not mentioning yet. etc. But some features are still not ported to V2, so it's taking me time and I want the top quality.

I've marked your email for pre-access, please follow up with me anytime.

[JJsilvera1]

Hello, I just signed up for the beta just now

[JJsilvera1]

@enricoros hello, I actually signed up but also I wanted to actually make a feature request that maybe most people don't recommend.

I just want to be able to have more than one window of big AGI open. It would be cool because then I can run multiple prompts and cross reference them in different tabs, but it seems like I get an error anytime I duplicate the tab that one sessions are already open

[enricoros]

I agree it would be great. It's in the backlog of features, and we'll get to it once we scale the operation. This will require a centralized system / server.

[JJsilvera1]

Hi @enricoros can I get access to v2 please?

[maxbraeutigam]

+1