One framework to evaluate any VLA model on any robot simulation benchmark.
Latest release v0.8.0 · 30 Sept 2026
2 files for Claude Code, Codex and OpenCode: vla-evaluation-harness CLAUDE.md, vla-evaluation-harness AGENTS.md — 1,307 tokens loaded in every session.
CLAUDE.md A 5 tok AGENTS.md A 1,302 tok These files are allenai/vla-evaluation-harness's own configuration — they tell Claude Code, Codex and OpenCode how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.