Benchmarks
Human preference benchmarks
People compare model answers and choose which they prefer.
We track 136 benchmarks in this topic. Main benchmarks are listed by how many models they cover, most first. Where a benchmark has separate results for a subset of its tasks, those results appear right under it. Click one for its leaderboard.
Showing 101–136 of 136 benchmarks.
- Arena ELO Vision - Captioning30 models
- Design Arena Elo (models/video)30 models
- Arena Hard22 models
- MM-MT-Bench13 models
- Arena ELO Video Edit12 models
- MT-Bench9 models
- AlpacaEval 2.01 model
- LMArena Text Leaderboard1 model