Tag: terminal-bench-science

Showing 1-1 of 1

Aug 28, 20268 min readAgent PlaybookIntermediate180 min build

Measure AI agent resolution rates with Terminal-Bench-Science v0.1 across 70 expert scientific workflows

Practical guide to running Terminal-Bench-Science v0.1: reproduce the 70-task benchmark, compute resolution rates (top model ≈30%), and build a small audit-ready evaluation pipeline.

terminal-bench-science benchmarks ai-agents scientific-workflows evaluation

+1 more

stanford