productivity benchmark
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.
Updated Aug 7, 2026
Published models3
Registry coverage3
MetricScore
EvidenceB
CoWorkBench leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | AC | 74.8% | 100.0% | 3 | C | |
| 2 | AC | 67.2% | 50.0% | 3 | C | |
| 3 | AC | 65.1% | 0.0% | 3 | C |
Score distribution
Top published rows on the benchmark's original scale.
CoWorkBench highlights
The top published results on this benchmark's own scale.
What is CoWorkBench?
Definition and scoring fields from the benchmark registry.
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.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- CoWorkBench
- Modality
- text
- Primary category
- productivity
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- coworkbench|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about CoWorkBench.
Which model scores highest on CoWorkBench?
Qwen3.8 Max is currently ranked first with 74.8%.
What does CoWorkBench measure?
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.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
3 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.