WeaveBench

What is WeaveBench?

WeaveBench evaluates computer-use agents on 114 long-horizon tasks across 8 real-world work domains. Each task interleaves GUI interaction with command-line…

Released

2026-06-08

Evaluates

General AI, Language & Knowledge, cs.AI

Openness

unknown

Importer

Claire Radar

Review status

source-reviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2606.09426

WeaveBench paper, code and dataset

• Paper https://arxiv.org/abs/2606.09426
• Code https://github.com/weavebench/WeaveBench
• Dataset https://huggingface.co/datasets/wanlilll/WeaveBench
• Website https://weavebench.github.io/

Which sources cite WeaveBench?

• WeaveBench

Related General AI, Language & Knowledge, cs.AI benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 3DCodeBench
• 5-Dialects-BN
• A Large-Scale Dataset and Benchmark
• A3S-Bench
• AARRI-Bench
• ABC-Bench
• ActBench