Updated continuously
The AI model leaderboard
The models that matter, ranked on the published benchmarks from Artificial Analysis. Intelligence, coding, Elo, speed and price, as measured, with nothing added by us.
Where these numbers come from
Every figure on this page is published by Artificial Analysis, who run the benchmarks. Text and coding models are ranked on their Intelligence Index and Coding Index. Image, video and speech models are ranked on Elo, scored from head-to-head human preference, shown with its 95% confidence interval because two models within the interval of each other are not meaningfully apart.
A dash means Artificial Analysis has not published that figure, not that the model scored zero. Nothing here is estimated or filled in by us.
Nothing measured in this category yet.
Nothing measured in this category yet.
Nothing measured in this category yet.
Nothing measured in this category yet.
Nothing measured in this category yet.
Not sure which of these fits your work?
The right model depends on the task, the language and what a wrong answer costs you. Fifteen minutes is usually enough to narrow it to one.
Book a free 15-minute call