Models used?

[matalab]

Is this library using the most recent Perplexity AI models? If not, how to define model(s) to be utilized? Asking because I noticed that I get different result than when using Perplexity AI web chat...

[matalab]

How are models listed in def _c(self, prompt: str, model: str) of labs.py related to models listed on ? Is it updated or needs to be updated with current models in Perplexity AI?

def _c(self, prompt: str, model: str) -> dict:
    assert self.finished, "already searching"
    assert model in [
        "mixtral-8x7b-instruct",
        "llava-7b-chat",
        "llama-2-70b-chat",
        "codellama-34b-instruct",
        "mistral-7b-instruct",
        "pplx-7b-chat",
        "pplx-70b-chat",
        "pplx-7b-online",
        "pplx-70b-online"
    ]
    self.finished = False
    self.history.append({"role": "user", "content": prompt, "priority": 0})
    self.ws.send("42[\"perplexity_playground\",{\"version\":\"2.1\",\"source\":\"default\",\"model\":\"" + model + "\",\"messages\":" + dumps(self.history) + "}]")

[matalab]

I have pruchased Pro subscription for Perplexity AI. Will this llibrary automatically utilize models available for Pro users or I need to specify it in python code? If it is the latter, how to do it?

[matalab]

As a PRO Perplexity user I can choose AI models listed in the user account settings control panel, where I can choose between: "Default", "Experimental", "GPT-4 Turbo", "Claude 3 Sonnet", "Claude 3 Opus", "Mistral Large". I have detected that "Claude 3 Opus" gives me best results. Can you please clarify, is there any relation between what I choose in user account settings and what model will utilize this python library? For example, for the official Perplexity API, it does not - there is no relation between user settings (and corresponding web app) and API models. Is it the same for this library or not?