Benchmark
SkillsBench
What is SkillsBench?
Measures whether curated Agent Skills raise agent pass rates, not raw model capability; the meaningful figure is the paired no-Skills versus curated-Skills…
- Released
- 2026-02-13
- Evaluates
- agent
- Openness
- unknown
- Importer
- Model reports
- Reported scores
- 1
Source provenance
- Original evidence https://skillsbench.ai/
SkillsBench links
- Website https://skillsbench.ai/
SkillsBench results and reported scores
Model reports
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Hy4 preview | Tencent | 62.9 | 2026-08-28 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.
Which sources cite SkillsBench?
- Hy4 preview Tencent, 2026-08-28