SkillsBench

Partners in Evaluation

Key Takeaways

• Grok 4.5 leads at 66.03%, ahead of GPT 5.5 with Codex at 62.55% and GPT 5.5 with OpenHands at 62.21%.
• Claude Opus 5 ranks sixth at 60.44% with skills, improving 15.25 points over its 45.19% score without skills.
• Skills helped every model we tested. The average score rose from 36.33% without them to 51.21% with them — a 14.88-point gain.
• The biggest gains came in the middle of the pack. MiniMax-M3 improved the most (+25.43 points), followed by DeepSeek V4 (+24.05) and Qwen 3.7 Plus (+23.15).
• Muse Spark 1.1 reached 59.19% at $0.45 per task, while Grok 4.5 led the benchmark at $0.58 per task.

Background

SkillsBench asks a focused question: do agents get better at software tasks when you hand them reusable, task-specific knowledge? To answer it, the benchmark runs each model twice with the same agent — once on its own, and once with relevant skills made available. The gap between the two runs is the value the skills add.

The benchmark was introduced by BenchFlow AI, who maintain an official leaderboard and the original repository. We partnered with them to integrate it into Valkyrie, which let us add it to the site quickly.

You can find that implementation in our new public benchmark registry, where the community can contribute benchmarks that are compatible with Valkyrie.

Results

Skills lifted strong and weak models alike, by an average of 14.88 points. The largest jumps came from MiniMax-M3 (+25.43 points), DeepSeek V4 (+24.05), Qwen 3.7 Plus (+23.15), Grok 4.3 (+21.33), and Claude Sonnet 4.6 (+21.22).

Grok 4.5 tops the with-skills board at 66.03%, followed by GPT 5.5 with Codex at 62.55% and GPT 5.5 with OpenHands at 62.21%. Grok 4.5 costs $0.58 per task; Muse Spark 1.1 reaches 59.19% at $0.45 per task.

Skill usage varies widely across models, and even across agents for the same model. On Codex, GPT 5.5 references skill paths far more than anything else — about 4.8 times per task — while the same base model on OpenHands, GPT 5.5, is among the lightest at roughly 1.2. After Codex, MiniMax-M3, GPT-5.6 Sol, and GPT-5.6 Terra reference skills most; Claude Sonnet 4.6 remains the lightest.

Methodology

Every model runs the same SkillsBench task set in two conditions:

• No Skills — the agent gets its default tools and prompt.
• With Skills — the same agent, plus task-specific skills.

The default agent is OpenHands-CLI, with the commit locked at 3ca17446c5d9c1e35e054803478a3501ec251ecf. The Codex row uses Codex with the same GPT-5.5 base model.

We run three trials per model in each condition. For each task we average the score across the three trials, then average those task scores into the model’s final accuracy.

The observed skill reference chart counts observed agent skill uses in With Skills runs. Each unique skill action counts once.

All runs use the public SkillsBench task definitions and run protocol.

@misc{li2026skillsbench,
title         = {SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks},
author        = {Xiangyi Li and Wenbo Chen and Yimin Liu and Shenghan Zheng and Xiaokun Chen and Yifeng He and Yubo Li and Bingran You and Haotian Shen and Jiankai Sun and Shuyi Wang and Qunhong Zeng and Di Wang and Xuandong Zhao and Yuanli Wang and Roey Ben Chaim and Zonglin Di and Yipeng Gao and Junwei He and Yizhuo He and Liqiang Jing and Luyang Kong and Xin Lan and Jiachen Li and Songlin Li and Yijiang Li and Yueqian Lin and Xinyi Liu and Xuanqing Liu and Haoran Lyu and Ze Ma and Bowei Wang and Runhui Wang and Tianyu Wang and Wengao Ye and Yue Zhang and Hanwen Xing and Yiqi Xue and Steven Dillmann and Han-chung Lee},
year          = {2026},
eprint        = {2602.12670},
archivePrefix = {arXiv},
url           = {https://arxiv.org/abs/2602.12670},
}