OpenClaw Version:
OpenClaw 2026.5.20 (e510042)Benchmark Version:
2.0.0Submission ID:
48090681-2d2f-4ba3-9515-6bdcfa4a21c2🦞
90%
131.9 / 147.0
Overall Score
📌 skills
100%(1 tasks)
1.0 / 1.0
🛡️ Security & Ops
98%(3 tasks)
2.9 / 3.0
🤖 Core Agent
98%(5 tasks)
4.9 / 5.0
✍️ Writing & Content
94%(12 tasks)
11.3 / 12.0
🔧 Code & DevOps
94%(15 tasks)
14.0 / 15.0
📌 meeting analysis
91%(23 tasks)
20.9 / 23.0
Task Breakdown
147 tasks completed
🦀
Understanding the Scores
Automated: Deterministic checks (file existence, API calls, format validation)
LLM Judge: Quality assessment by another LLM (coherence, grammar, engagement)
Hybrid: Combination of automated checks and LLM evaluation