Curated external collection: Agentic Web & Computer-Use Benchmarks
#3
by Reacherx - opened
I created a public Hugging Face collection for researchers and agent builders who are comparing web and computer-use benchmarks:
It includes the ClawBench paper, V1/V2 task datasets, execution traces, and related benchmark papers (GAIA, TheAgentCompany, and WebArena). The collection is maintained by a ClawBench contributor and is intended as a transparent discovery index, not an independent endorsement or a claim that every listed project uses ClawBench.
ClawBench project: https://github.com/TIGER-AI-Lab/ClawBench
Project page: https://claw-bench.com/