agents benchmark
SkillsBench Leaderboard
SkillsBench evaluates coding agents on self-contained programming tasks across diverse software development scenarios.
Updated Sep 5, 2026
Models9
Model coverage9
MetricScore
EvidenceB
SkillsBench Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAC | Score70.20% | Percentile100.00% | Participants9 | EvidenceC | Evaluated |
| Rank02 | ModelTE | Score62.90% | Percentile87.50% | Participants9 | EvidenceC | Evaluated |
| Rank03 | ModelAC | Score59.20% | Percentile75.00% | Participants9 | EvidenceC | Evaluated |
| Rank04 | ModelTE | Score55.30% | Percentile62.50% | Participants9 | EvidenceC | Evaluated |
| Rank05 | ModelAC | Score54.90% | Percentile50.00% | Participants9 | EvidenceC | Evaluated |
| Rank06 | ModelAC | Score48.20% | Percentile37.50% | Participants9 | EvidenceC | Evaluated |
| Rank07 | ModelAC | Score45.70% | Percentile25.00% | Participants9 | EvidenceC | Evaluated |
| Rank08 | ModelME | Score44.30% | Percentile12.50% | Participants9 | EvidenceC | Evaluated |
| Rank09 | ModelAC | Score28.70% | Percentile0.00% | Participants9 | EvidenceC | Evaluated |
SkillsBench Highlights
The leading models and scores on this benchmark.
SkillsBench Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for SkillsBench
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is SkillsBench?
What SkillsBench measures and how its scores work.
SkillsBench is a benchmark for evaluating coding agents on self-contained programming tasks.
It measures practical engineering skills across diverse software development scenarios, using the Score metric in ratio units.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- SkillsBench
- Modality
- text
- Primary category
- agents
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about SkillsBench.
Which model scores highest on SkillsBench?
Qwen3.8 Max is currently ranked first with 70.20%.
What are the top three models on SkillsBench?
The current leaders are Qwen3.8 Max (70.20%), Hy4 preview (62.90%), and Qwen3.7 Max (59.20%).
Which SkillsBench model has the lowest official input price?
Qwen3.6-35B-A3B has the lowest matched official input price at $0.25 input / $1.5 output per 1M tokens.
Which models are fastest among SkillsBench results?
The fastest matched records are Qwen3.8 Max (60.87 tok/s via DeepInfra).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.