Model Lab · R&D
Two models, one prompt — you pick the winner
Blind head-to-head between our AI providers. Pick the better output; the team's preference compounds into a default we pin.
Head-to-head prompt
Two models answer the same prompt, blind. You pick the better one; we learn which to pin.
Running win-rate
Across all judged comparisons — pins the team's preference over time
Anthropic0 wins · 0%
OpenAI0 wins · 0%
Ties: 0