CoWorkBench
Qwen3.8-Omni-Flash 75.3% leads since Sep 2026 In use Agent completes long-horizon productivity tasks across computer science, finance, law and medicine. Qwen in-house scoring (method unpublished).
Lab-created in-house eval (Alibaba Qwen). Only Qwen reports scores. Tasks and graders are not public.
- Level
- Expert Specialists with years of training, often over hours of work
- Scored
- Graded by a model A language model scores each response, usually against a rubric.
- Answers
- Open-ended Free-form text, documents, designs or media judged as a whole.
Qwen3.8-Omni-Flash leads with 75.3%
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
- 1 Qwen3.8-Omni-Flash 75.3%
- 2 Qwen 3.8 Max 74.8%
- 3 Qwen3.8 Flash 73.9%
- 4 Qwen 3.8 27B 70.7%
- 5 Qwen3.7-Max 67.2%
- 6 Qwen3.7-Plus 65.1%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Qwen3.8-Omni-Flash | 75.3% | Sep 2026 | Model card | |
| 2 | Qwen 3.8 Max | 74.8% | Aug 2026 | Model card | |
| 3 | Qwen3.8 Flash | 73.9% | Aug 2026 | Model card | |
| 4 | Qwen 3.8 27B | 70.7% | Aug 2026 | Model card | |
| 5 | Qwen3.7-Max | 67.2% | May 2026 | Model card | |
| 6 | Qwen3.7-Plus | 65.1% | Jun 2026 | Model card |
3 models have held number one since 2023
Lab claims count from the day the card came out, independent runs from the model’s release.
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
One lab put it in a launch post in the last six months
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| 18 Sep 2026 | Qwen3.8-Omni-Flash | 75.3 | CoWorkBench Long-horizon office work |
| 26 Aug 2026 | Qwen3.8 Flash | 73.9 | Long-horizon office work CoWorkBench |
| 14 Aug 2026 | Qwen 3.8 27B | 70.7 | Long-horizon office work CoWorkBench |
| 3 Aug 2026 | Qwen 3.8 Max | 74.8 | CoWorkBench |
| 1 Jun 2026 | Qwen3.7-Plus | 65.1 | CoWorkBench |
| 20 May 2026 | Qwen3.7-Max | 67.2 | CoWorkBench |
One system card reports it
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| 22 Sep 2026 | Qwen3.8-Omni-Flash technical report | 75.3 | CoWorkBench, in-house long-horizon office tasks (Table 2) | Capabilities |
More in Economic work
- JobBench In use 130 tasks from 35 occupations workers most want delegated. Agent works through messy reference files, graded by fact-anchored binary rubrics. GLM-5.3 61.4%
- GDPval In use Real work tasks from 44 occupations with reference files. Experts blindly compare model deliverables with expert work. Metric is win-or-tie rate.
Seed2.1 Pro 87.9% - GDPval-AA v2.1 In use Agents with shell and web access complete 220 GDPval gold tasks from 44 occupations. Blind pairwise comparisons give Elo ratings. Claude Opus 5.5 1,846
- AA-Briefcase v1.1 In use Version 1.1 of Artificial Analysis's multi-hour office work evaluation, with Elo ratings fitted by Crowd-BT. Claude Opus 5.5 1,822
- OfficeQA In use Grounded reasoning over a large corpus of U.S. Treasury Bulletin documents: locate tables and compute precise numeric answers. Claude Opus 4.7 86.3%
- SpreadsheetBench In use Real spreadsheet manipulation questions from Excel forums. Output workbook is checked cell-by-cell against reference solutions.
MiniMax-M3 89.3%