Perplexity introduces Model Council, a safer and more efficient way to get answers by querying several AI models at once. Published February 5, 2026, the service gathers outputs from different models, compares them, and gives you a synthesized response that highlights agreements and disagreements.

What is Model Council

Model Council is a multi-model research feature: instead of testing models one by one, you make a single query and the system runs it across several models at the same time. Then a synthesizer model reviews the answers, tries to resolve contradictions when possible, and returns a consolidated reply with clear signals about where sources agree or differ.

Why switch to this flow? Think about asking three people the same question—wouldn’t you want a quick summary of where they line up and where they don’t, rather than reading three separate long replies?

For example, a query can run simultaneously on Claude Opus 4.6, GPT 5.2 and , and the interface will show the final synthesis along with notes about points of disagreement.