SkillsBench leaderboard
Also searched as: SkillsBench (with skills) · Vals AI
As of 6 Nov 2026, the top score on SkillsBench (also called SkillsBench (with skills)) is 69.8% by DeepSeek V4.1 Flash (DeepSeek), ahead of Grok 4.5 at 66.0%, across 34 model configurations with a published result (Vals AI, result dated 29 Oct 2026). Agents complete tasks with and without a library of reusable skills; the score reflects task success with skills available. SkillsBench is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 6 Nov 2026)
- DeepSeek V4.1 Flash holds the top SkillsBench score at 69.8% (Vals AI, 29 Oct 2026).
- Grok 4.5 is second on SkillsBench at 66.0% (Vals AI, 29 Oct 2026).
- SkillsBench: How much reusable skills help an agent complete tasks. Run by Vals AI (Vals AI).
- 34 model configurations have a published SkillsBench result on BenchLeader (BenchLeader, 6 Nov 2026).
- SkillsBench does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
How much reusable skills help an agent complete tasks. Run by Vals AI.
- Published by
- Vals AI
- Category
- Agents & tools
- Index weight
- Reference only
- Models
- 34
- Data as of
- 6 Nov 2026
Vals AI (vals.ai).
What the test looks like
Agents complete tasks with and without a library of reusable skills; the score reflects task success with skills available.
How it is scored
Percent of tasks completed, run by Vals AI.
What to keep in mind
Measures a specific agent design choice; not a general capability test.
- 1DeepSeek V4.1 Flash (high)69.8%
- 2Grok 4.5 (high)66.0%
- 3Gemini 3.7 Flash (high)65.9%
- 4GPT 5.5 Codex62.6%
- 5GPT-5.5 (xhigh)62.2%
- 6Claude Fable 5.161.5%
- 7GPT-5.6 Luna (max)60.5%
- 8Claude Opus 560.4%
- 9Claude Opus 4.859.2%
- 10Muse Spark 1.1 (xhigh)59.2%
- 11GPT-5.6 Terra (max)58.9%
- 12Gemini 3.8 Flash (high)58.0%
- 13Grok 4.6 (high)55.8%
- 14Qwen3.7 Plus54.3%
- 15GPT-5.6 Sol (max)54.1%
| # | |||||
|---|---|---|---|---|---|
| 1 |  DeepSeek V4.1 Flashhighopen ↗ | 69.8% | 55.7 | $0.085 | 2026-10-29 |
| 2 |  Grok 4.5high | 66.0% | 56.9 | $0.578 | 2026-10-29 |
| 3 |  Gemini 3.7 Flashhigh | 65.9% | 62.2 | $1.80 | 2026-10-29 |
| 4 |  GPT 5.5 Codex | 62.6% | – | $1.57 | 2026-10-29 |
| 5 | GPT-5.5xhigh | 62.2% | 64.3 | $2.54 | 2026-10-29 |
| 6 |  Claude Fable 5.1 | 61.5% | 66.4 | $4.52 | 2026-10-29 |
| 7 | GPT-5.6 Lunamax | 60.5% | 56.3 | $0.207 | 2026-10-29 |
| 8 | Claude Opus 5 | 60.4% | 65.2 | $2.51 | 2026-10-29 |
| 9 | Claude Opus 4.8 | 59.2% | 60.0 | $4.77 | 2026-10-29 |
| 10 |  Muse Spark 1.1xhigh | 59.2% | 59.5 | $0.449 | 2026-10-29 |
| 11 | GPT-5.6 Terramax | 58.9% | 60.7 | $1.75 | 2026-10-29 |
| 12 | Gemini 3.8 Flashhigh | 58.0% | 62.0 | $2.34 | 2026-10-29 |
| 13 | Grok 4.6high | 55.8% | 60.5 | $0.921 | 2026-10-29 |
| 14 |  Qwen3.7 Plus | 54.3% | 57.7 | $0.137 | 2026-10-29 |
| 15 | GPT-5.6 Solmax | 54.1% | 65.4 | $4.14 | 2026-10-29 |
| 16 | DeepSeek V4 Promaxopen ↗ | 53.8% | 57.0 | $0.375 | 2026-10-29 |
| 17 | Muse Spark 1.2xhigh | 53.0% | 61.7 | $0.649 | 2026-10-29 |
| 18 | Gemini 3.5 Flashhigh | 52.7% | 62.5 | $1.25 | 2026-10-29 |
| 19 | GPT-5.4xhigh | 51.7% | 63.2 | $2.04 | 2026-10-29 |
| 20 |  MiniMax M3open ↗ | 51.5% | 56.4 | $0.505 | 2026-10-29 |
| 21 | DeepSeek V4 Flashhighopen ↗ | 50.7% | 57.3 | $0.126 | 2026-10-29 |
| 22 |  Kimi K2.7 Codeopen ↗ | 50.0% | 54.2 | $0.443 | 2026-10-29 |
| 23 | Claude Sonnet 4.6 | 49.0% | 58.2 | $1.38 | 2026-10-29 |
| 24 |  GLM 5.3maxopen ↗ | 47.5% | 59.0 | $0.714 | 2026-10-29 |
| 25 | Claude Sonnet 5 | 46.5% | 55.9 | $2.90 | 2026-10-29 |
| 26 | GLM 5.2maxopen ↗ | 45.1% | 59.5 | $0.811 | 2026-10-29 |
| 27 | Qwen3.8 Maxmax | 42.0% | 58.8 | $0.525 | 2026-10-29 |
| 28 | Grok 4.3high | 40.6% | 56.6 | $0.420 | 2026-10-29 |
| 29 | GLM 5.3 Flashmaxopen ↗ | 40.2% | 54.2 | $0.078 | 2026-10-29 |
| 30 | Qwen3.8 27Bxhighopen ↗ | 38.1% | 52.5 | $0.818 | 2026-10-29 |
| 31 | Inkling Smallopen ↗ | 33.6% | 52.1 | $0.148 | 2026-10-29 |
| 32 |  Ling 3.0 Flashopen ↗ | 27.9% | 51.8 | $0.061 | 2026-10-29 |
| 33 | Inklingopen ↗ | 26.1% | 47.5 | $0.827 | 2026-10-29 |
| 34 |  Mercury 2.5high | 18.1% | – | $0.101 | 2026-10-29 |
More agents & tools benchmarks · All LLM benchmarks
Related benchmarks
Top head-to-heads on SkillsBench
Cite as: BenchLeader, “SkillsBench leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/vals_skillsbench.
SkillsBench: questions
- What does SkillsBench measure?
- Agents complete tasks with and without a library of reusable skills; the score reflects task success with skills available. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on SkillsBench?
- As of 6 Nov 2026, the highest score on SkillsBench is 69.8% by DeepSeek V4.1 Flash (result dated 29 Oct 2026), ahead of Grok 4.5 at 66.0%. Scores come from Vals AI and are re-read daily.
- Which AI model leads SkillsBench?
- DeepSeek V4.1 Flash leads SkillsBench with 69.8% as of 6 Nov 2026 (result dated 29 Oct 2026), ahead of Grok 4.5 at 66.0%.
- How many models have SkillsBench results?
- 34 model configurations have a SkillsBench result on BenchLeader, all taken from Vals AI.
- Who runs SkillsBench and how often is it updated?
- SkillsBench is published by Vals AI. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does SkillsBench count toward the BenchLeader Index?
- No. SkillsBench is shown for reference but left out of the composite index.