Benchmark run
90.3TRUNK
Code
89.5
Repo fix
100.0
Agent
100.0
Tool calls
100.0
Reason
83.3
Long docs
95.0
Instructions
48.9
2026-10-01testpack 4.0.0-alpha.3cli 0.4.271 tasks · v4 · scored locally
How does your rig stack up?
Same 71 tasks, scored on your machine. No API key, no account. One command:
$ npx @pipelinescore/cli