all the models — AI benchmark observatory
← Benchmarks

CoWorkBench

CoWorkBench is Qwen's internal cowork benchmark for evaluating long-horizon office and productivity agent tasks across domains such as computer science, finance, law, and medicine.

id coworkbench · max 1 · 6 models reported

#ModelScore
1Qwen3.8 Max
Alibaba Cloud / Qwen Team · open
0.75
2Qwen3.8 Flash
Alibaba Cloud / Qwen Team
0.74
3Qwen3.8-Flash-Next
Alibaba Cloud / Qwen Team · open
0.74
4Qwen3.8-27B
Alibaba Cloud / Qwen Team · open
0.71
CoWorkBench Leaderboard · all the models