Other · model
UI-Mate-27B benchmark scores
As of 2026-11-22, UI-Mate-27B (Other) has measured scores on 1 benchmark and estimated scores on 15 more.
| Benchmark | Score | Source |
|---|---|---|
| GDPval-AA | 38.0% | estimated ± 8.5 pp, medium confidence |
| BrowseComp | 83.8% | estimated ± 4.7 pp, high confidence |
| APEX-Agents-AA | 39.0% | estimated ± 7.8 pp, medium confidence |
| DeepSearchQA | 83.6% | estimated ± 7.9 pp, low confidence |
| CyberGym | 75.6% | estimated ± 11.3 pp, low confidence |
| JobBench | 46.2% | estimated ± 11.1 pp, low confidence |
| WideResearch | 81.4% | estimated ± 2.1 pp, medium confidence |
| AA Agentic Index | 28.3% | estimated ± 8.0 pp, medium confidence |
| Gert Labs | 48.9% | estimated ± 10.9 pp, low confidence |
| ApprenticeBench | 12.4% | estimated ± 6.3 pp, low confidence |
| OSWorld-Verified | 77.0% | measured |
| AA ITBench | 42.4% | estimated ± 4.7 pp, low confidence |
| OSWorld 2.0 | 11.4% | estimated ± 3.6 pp, high confidence |
| Claw-Eval | 67.4% | estimated ± 6.2 pp, low confidence |
| Toolathlon | 55.3% | estimated ± 9.3 pp, medium confidence |
| ResearchClawBench | 18.6% | estimated ± 2.1 pp, medium confidence |