Artificial Analysis tool to create custom benchmarks for any use case
Turn your own tasks into a repeatable benchmark, run frontier models on it, and compare quality against cost per task. Bring your own dataset, agent trajectories, or your own agent.

Reactions (1)
Similar products
by description similarity
BenchFlow
github.comrun AI benchmarks as an API
DesignArena
designarena.aicrowdsourced benchmark for AI-generated UI/UX
Benchgraph
news.ycombinator.comRun graph database benchmark on your own dataset
A tool to benchmark LLM APIs (OpenAI
llmapitest.comClaude, local/self-hosted)
Τ³-Bench is out
news.ycombinator.comcan agents handle complex docs and live calls?