Two AI models. One question. You decide who answered better. 25 questions across 5 topics. Every vote shapes the verdict.
One vote per session · A/B positions are randomized each round · Identities revealed when voting closes
Two AI models — Claude by Anthropic and ChatGPT by OpenAI. We asked them the same 25 questions across 5 topics. Same prompt. Same constraints. No coaching.
You decide who answered better. One question goes live each day for 72 hours. Vote before the window closes — and you won't know which model wrote which response until after you have.
Why blind? Brand loyalty is loud. Stripping the logos is the only way to find out which answer people actually prefer.