Perplexity launches Model Council to compare AI models
Perplexity has launched Model Council, a feature that queries three AI models and combines their responses into one. It is already available on the web for Perplexity Max subscribers, while the mobile version is coming soon.

Perplexity has launched Model Council, a feature that queries several artificial intelligence models at the same time and combines their responses into one. The tool is designed to save you from manually testing different models whenever you need a reliable answer.
The company says model performance varies depending on the task. One model may be better at coding, while another delivers better results when researching, analyzing information, or generating ideas. That is why choosing the right model is not always obvious before you ask a question.
How Model Council works
When you select Model Council in the web version of Perplexity, the same question is sent to three models available on the platform. The examples mentioned by the company include Claude Opus 4.6, GPT 5.2, and Gemini 3.0.
A synthesizer model then takes over. It reviews the responses, tries to resolve contradictions, and presents a combined answer. It also points out where the models agree and where they differ.
The idea is simple: if several responses reach a similar conclusion, you can move forward with more confidence. If they disagree, you get a clear signal that you should review the data or investigate further before making a decision.
Every AI model has blind spots.
What you can use it for
Model Council is designed for queries where an error, omission, or biased perspective could have serious consequences. Perplexity highlights these use cases:
- Financial research: comparing analysis of stocks, markets, or investment decisions.
- Complex decisions: evaluating a major purchase, career change, or business strategy from several perspectives.
- Brainstorming: combining different ways of generating ideas for trips, gifts, or content.
- Verification: cross-checking information when you need more confidence before accepting it as valid.
This does not mean the combined answer is automatically correct. The system still depends on the models it consults and the information available to them. Its value lies in making agreements and disagreements visible, something you would normally discover only by switching between models yourself.
Who can use it
Model Council is already available on the web to subscribers of the Perplexity Max plan. Compatibility with mobile apps is coming soon, according to the company.
The feature reflects a practical shift in how people use AI assistants. Instead of looking for a single model that can handle everything, Perplexity is betting on combining several models and showing when their answers align or contradict one another. The key question will be whether this comparison actually improves accuracy in real-world queries, especially when mistakes come at a high cost.