Artificial Analysis independently benchmarks AI agents and models, the cloud infrastructure that serves them, and the chips they run on. Its data helps developers and companies compare intelligence, speed, price, latency, and other performance characteristics across the AI stack.
Products & media
Independent benchmarks compare AI models, agents, inference providers, and chips across quality, speed, latency, price, and openness.
AA-3D is Artificial Analysis's new benchmark that tests AI models as agents building interactive 3D scenes. Models receive written briefs, work in an offline sandbox with Blender, physics, and a browser, then produce three.js web pages that a human panel judges in blind pairwise comparisons to produce Elo rankings.
A developer API for structured model metadata, benchmark results, pricing, and performance data across language, image, video, music, and speech.
A custom benchmarking platform that evaluates AI models on an organization’s files, agent traces, or coding environment across quality, cost, and speed.
AA-Robotics v0.1 is Artificial Analysis’s research preview for measuring how well frontier language models understand, plan, and act in the physical world. Each model is given control of a 6-axis robot arm plus camera feeds and must write code to complete dexterity, spatial-reasoning, and adaptability tasks zero-shot.
Team
Founders
2 people
Jobs
See all jobs for Artificial Analysis here.
