pinchbench

Run PinchBench benchmarks to evaluate OpenClaw agent performance across real-world tasks. Use when testing model capabilities, comparing models, submitting benchmark results to the leaderboard, or checking how well your OpenClaw setup handles calendar, email, research, coding, and multi-step workflows.

Try your first task

Run a small PinchBench sanity benchmark locally

Start a fresh conversation.

Use the pinchbench skill to run only `task_sanity` with `--no-upload` from <skill_directory>, using my provider-prefixed model ID <provider/model>. Report the locally saved result and do not register a token or upload anything.

Required inputs

<skill_directory>: path to the cloned PinchBench skill directory
<provider/model>: provider-prefixed model identifier
Python 3.10+, uv, and a running OpenClaw instance

Suggested template by Skill Leaderboard, grounded in the linked source. Task evidence ↗

Inside SKILL.md

123 lines

Section content has not been verified for this skill yet.

Full source remains withheld under the current source-display policy. Read the author’s source ↗