Compare models
Pick 2 to 4 models to see their scores, prices, speed and specs side by side. The best value in each row is highlighted.
Loading data…
Comparison
Highlighted cells are the best value in the row (highest score, speed and context; lowest price and latency). n/a = no data published. Benchmarks are % correct. TTFT includes reasoning time: models that hide their reasoning at high effort can take minutes before the first token.