By Manisha Sharma
Publication Date: 2026-02-11 12:24:00
As AI models become more capable and more specialised, the question for users is no longer whether to use AI, but which AI to trust for a given task.
Perplexity’s latest product update, called Model Council, is an attempt to address that problem directly. Rather than asking users to switch between models and compare answers on their own, the company is now offering a single interface where multiple models respond to the same query at once.
The feature reflects a growing reality in frontier AI: performance varies widely depending on the task, and no single model consistently delivers the best answer across research, coding, decision-making and creative work.
Turning Model Selection Into a System Feature
Model Council is positioned as a research-focused mode within Perplexity’s main interface. When enabled, a user’s query runs simultaneously across three different AI models available on the platform.
Examples cited include Claude Opus 4.6, GPT 5.2, and Gemini 3.0, though the exact mix may vary. A separate synthesiser model then reviews the responses, resolves conflicts where possible, and produces a single answer.
Crucially, the output does not flatten differences entirely. Instead, it highlights where models agree and where their answers diverge—giving users signals about confidence, uncertainty and potential blind spots.
This approach shifts model comparison from a manual exercise into an embedded workflow, particularly for users who rely on AI outputs for decisions…