Introducing Model Council in Perplexity. Run three frontier models at once, compare outputs, and get a more accurate, higher‑confidence answer. Available now on web only for Perplexity Max subscribers.

Model Council runs your query through three selected AI models in parallel and compares the responses. A separate model then reviews and synthesizes the results, highlighting where they agree, disagree, and what each one found.

View results side by side, get clear insight into the evidence and disagreements, and see what each model uniquely contributes to the final answer.

Learn more about Model Council on our blog:

me before for every decision: "hold on let me ask the Model Council"

@perplexity_ai dude, Agentspark literally has the same thing for the lowest paid tier users and it uses all available models at once. And yet you’re limiting yours to max users and not even using all of the models. What’s the point?

Where are our 200 pro searches a week or 50-100 labs?, Bring that back First! ❗

This is awesome. More of this please.

This is so sick!! Been wanting this for a long time. Hopefully it’ll make its way to Pro soon.

Like ???

Cool concept could be very helpful for people who dont know which model to use for their particular problem

Running multiple frontier models side by side is a smart move. Better comparisons, better answers 🔥

This is great! Perplexity gets better and better.

Is this coming to Pro subscribers anytime soon? Love this! Also iOS apps too?

If the 3 smartest models on earth agree, the confidence interval just hit the ceiling. 🧠🔥

Love this, I do it manually but this is so much more convenient

@karpathy LLM council was the council i played around. Changed that C-Suite council and connected with MCPs like salesforce, Quickbooks, and other essentials. Good results.

Woah Nice ship guys 🤩

Yes, my SEELE dream draws closer

This is a clever approach to enhancing AI reliability. Triangulating outputs from multiple models should definitely reduce hallucination and boost confidence in complex queries. Eager to see how this performs.

It’s an amazing thing and I was wondering when it would actually arrive, I would like two things, in order of importance: separate chats in the spaces started among themselves improvement of text-to-speech Perplexity is truly an amazing service.

go fuck yourselves you are taking everything away from those who pay for the pro you have reduced the limits for deep research from 600 to 20 you have removed the image I hope that you get a fine I pay 20 $ to get nothing and below in the comments all ass kissers with verified

this is an amazing idea

Roll out a "light" model council version to Pro users?!? 🙏

This would be a killer feature in Pro, as it saves time by automating something many people already do. However, it adds nothing new and provides no additional access (e.g., limited uses of GPT 5.2 Pro), so the Max paywall is unfortunate and a lost opportunity.

🧑‍⚖️👩‍⚖️👨‍⚖️

I love that you synthesize the answers and present where they agree/disagree. Much more useful than simply running the same prompt through several AI models (which of course I already do).

In my enterprise max it is not even getting selected, you people looted money from me with this advertisement 😞

@AravSrinivas This is badass

we want Comet on iOS... any news on that?

@arpitrage re your earlier tweet

This is the future

You all screwed the Pro subscription with your new underhanded restrictions like Deep Research going from 200 per month down to just 20 per month, and limiting file retention from 30 days to 7, plus morewithout even an annoucement. I've canceled my subscription and won't be back.

this is sweet! this is what I'm doing on a daily basis anyways. Yo @perplexity_ai - can you hit me with one of your sweet discount codes?! :)

Burn your credits fast, now 3x more.

i’ve done this manually for a few years now, pretty nice feature to add.

Pro ?

very cool but we need it in pro too please