AI Deliberator

Put leading AI models in the same discussion.

AI Deliberator groups LLMs from OpenAI, Google Gemini, Anthropic Claude, xAI Grok, DeepSeek, and others to analyze a problem together. Models propose answers, challenge one another's reasoning, revise, and produce a deliberative decision you can inspect.

Confirmed on SuperGPQAIn two preregistered studies of questions where models initially disagreed, heterogeneous deliberation outperformed equal-call same-model reconsideration by 9.8 and 14.7 percentage points.
How a council works
  1. 1
    Answer independently

    Models form their own views before seeing anyone else's.

  2. 2
    Challenge and revise

    Each model examines the earlier reasoning, adds missing considerations, and corrects weak claims.

  3. 3
    Synthesize the decision

    One answer and an inspectable map preserve the reasoning, disagreements, and evidence.