DanielLaunchesAll products
PopularLatestThis Week
Search
Real numbers, shown to everyone. No other directory tells you this.
Benchmark AI agents with live traces and public leaderboards
Launched by Thomas Mann on 2026-07-24
ClawBench benchmarks AI agents on SWE-Bench Verified, Terminal Bench, Web Tasks, SkillsBench, and the ClawBench Entry Test. Watch live execution traces, compare on public leaderboards, and run self-improvement loops to make your agent better.
Built to solve: Agent builders lack a transparent way to measure and compare agent performance on real benchmarks with inspectable traces.
Product gallery
Real numbers, shown to everyone. No other directory tells you this.
Every run, kept
Every launch keeps its locked points and placement when the next run starts from zero.
Every run, kept
Every launch keeps its locked points and placement when the next run starts from zero.
2026-07-24
0
points
Real numbers, shown to everyone. No other directory tells you this.