Claw-SWE-Bench Model × Claws OpenClaw × Models About Paper GitHub
Agent-harness benchmark · 350 tasks · 8 languages

Claw-SWE-Bench Leaderboard

View leaderboard Read the paper
Full-350 · cost vs. quality

Cost–Resolve Pareto

Total API cost (log scale) vs Pass@1 for all 7 harnesses × 3 models using means over three runs. The line connects non-dominated pairs.

image/svg+xml Matplotlib v3.10.8, https://matplotlib.org/ $10 $20 $50 $100 $200 $500 Total API cost for full 350-instance run (USD, log scale) 40 50 60 70 80 Resolved rate / Pass@1 (%) nanobot GenericAgent Hermes Agent Meta-Harness GenericAgent Pareto frontier GLM 5.1 Qwen 3.6-flash DeepSeek-V4.1-Flash OpenClaw Hermes Agent nanobot ZeroClaw GenericAgent DeepSeek Harness Meta-Harness