SkillsBench
SkillsBench results reported for 4 models across 2 versions. Updated 2026-09-24.
Report an issueVersions
2 versions · Metric: Task success rate
| Version | Released | Models | Results |
|---|---|---|---|
| SkillsBench 78-task self-contained subset — OpenCodeConfiguration: OpenCode harness | April 2, 2026 | 3 | 3 |
| SkillsBench 86-task set — With skillsConfiguration: With skills | February 10, 2026 | 1 | 1 |
Recently reported results: SkillsBench 78-task self-contained subset — OpenCode, OpenCode harness
Most recent configuration: SkillsBench 78-task self-contained subset — OpenCode
| Model | Score | Source |
|---|---|---|
| Qwen3.7-Plus | 54.9% | alibabacloud.com for Qwen3.7-Plus on SkillsBench 78-task self-contained subset — OpenCode (opens in a new tab)Task success rateReported Launch postAlibaba Cloud CommunitySelf-reportedQwen TeamRecord for Qwen3.7-Plus on SkillsBench 78-task self-contained subset — OpenCode |
| Qwen3.7-Max | 59.2% | alibabacloud.com for Qwen3.7-Max on SkillsBench 78-task self-contained subset — OpenCode (opens in a new tab)Task success rateReported Launch postAlibaba Cloud CommunitySelf-reportedQwen TeamRecord for Qwen3.7-Max on SkillsBench 78-task self-contained subset — OpenCode |
| Qwen3.6-Plus | 45.7% | alibabacloud.com for Qwen3.6-Plus on SkillsBench 78-task self-contained subset — OpenCode (opens in a new tab)Task success rateReported Launch postAlibaba Cloud CommunitySelf-reportedQwen TeamRecord for Qwen3.6-Plus on SkillsBench 78-task self-contained subset — OpenCode |