The benchmark of AI.
Get in touchArtificial Analysis is the independent benchmarking company for AI. We measure the agents and models, the clouds that serve them, and the chips they run on, helping developers and companies optimize the entire AI stack.
Publicly referenced by the leading AI labs, publications, and institutions
What we measure
Our methodologyAgents
Models paired with a harness (agent loop, tools) to complete multi-step tasks.
Models
Language, image, video, speech, and music models, both proprietary and open weights.
Cloud
The infrastructure that hosts and serves models, including serverless and dedicated offerings.
Chips
The accelerators that models run on: NVIDIA and AMD GPUs, Google TPUs, AWS Trainium, Cerebras WSE, SambaNova RDU and more.
Benchmarks at scale
500+
Models benchmarked
100+
Inference providers
1,000+
Endpoints
1T+
Evaluation tokens
Founders
