DEV Community

Nazar Batrak
Nazar Batrak

Posted on

Why I stopped trusting a single AI model

Ask Claude, GPT, and Gemini the same question and you get three different answers — and
you're left guessing which to trust.

So I built PolyHelper: send one prompt, it runs across many models at once and shows where
they agree, where they disagree, and a consensus — with an audit trail behind it.

Three modes:

  • Convergent — one answer + how strongly the models agreed
  • Divergent — the perspectives side by side
  • Artifact council — models critique and rank each other

Open source, Apache-2.0: https://github.com/PolyHelper/polyhelper

What consensus method would you actually trust? Majority vote feels naive, LLM-as-judge
has biases. How would you approach it?

Top comments (0)