Evaluates AI agents on realistic workplace tasks in a simulated corporate environment — email, calendar, project management, code review, and cross-team coordination.
This definition has not completed editorial and provenance review. Verify its metric, version, source, and evaluation configuration upstream before using it for a model-selection decision.
Tests holistic agentic workplace capabilities, not just isolated tool use. Highly relevant to AI-assisted knowledge work.
Non-standard score scale (0.28–6.78). Relatively new benchmark.
Higher is better. Performance score (domain-specific rubric).
Non-standard score scale (0.28–6.78).
| # | Model | Score |
|---|---|---|
| 1 | 59.6 | |
| 2 | 59.3 | |
| 3 | 56.9 | |
| 4 | 56.8 | |
| 5 | 56.4 | |
| 6 | 52.7 | |
| 7 | 40.3 | |
| 8 | 40.2 | |
| 9 | 39.9 | |
| 10 | 37.2 | |
| 11 | 33.7 | |
| 12 | 32.8 | |
| 13 | 32.8 | |
| 14 | 31.9 | |
| 15 | 31.6 | |
| 16 | 24.2 | |
| 17 | 23.9 | |
| 18 | 23.8 | |
| 19 | 18.3 | |
| 20 | 17.6 | |
| 21 | 17.4 | |
| 22 | 15.3 | |
| 23 | 15.2 | |
| 24 | 14.8 | |
| 25 | 6.9 | |
| 26 | 6.8 | |
| 27 | 6.8 | |
| 28 | 6.8 | |
| 29 | 6.6 | |
| 30 | 5.2 | |
| 31 | 4.9 | |
| 32 | 4.8 | |
| 33 | 4.6 | |
| 34 | 4.2 | |
| 35 | 4.2 | |
| 36 | 4.0 | |
| 37 | 4.0 | |
| 38 | 3.9 | |
| 39 | 3.9 | |
| 40 | 3.7 | |
| 41 | 3.6 | |
| 42 | 3.6 | |
| 43 | 3.5 | |
| 44 | 3.5 | |
| 45 | 3.5 | |
| 46 | 3.4 | |
| 47 | 3.4 | |
| 48 | 3.4 | |
| 49 | 3.4 | |
| 50 | 3.4 |