A.S.E (AICGSecEval): leaderboard
Repository-level AI-generated code security benchmark from Tencent Wukong Code Security Team (arXiv:2508.18106v3, Tables 2-3) evaluating 37 Code LLMs on real GitHub CVE patches with dual mutations. Scored by weighted Overall score (0.6*Security + 0.3*Quality + 0.1*Stability) on a 0-100 scale.
Metric: Overall weighted score (0-100). Source: arxiv.org. Saturation forecast: Not forecast. 33 models tracked. Updated 21 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.7 Sonnet (20250219) | 63.01 | #421 |
| 2 | Qwen 3 235B A22B 2507 Instruct | 60.15 | #616 |
| 3 | Qwen 3 Coder 480B A35B Instruct | 59.31 | #652 |
| 4 | DeepSeek V3 (0324) | 58.59 | #715 |
| 5 | Claude Sonnet 4 (20250514) | 57.14 | #405 |
| 6 | Qwen 3 30B A3B 2507 Instruct | 56.59 | #1100 |
| 7 | Kimi K2 (Preview) | 55.29 | #307 |
| 8 | GPT-4o (2024-11-20) | 55.1 | #859 |
| 9 | Qwen-Coder-Plus-20241106 | 53.55 | |
| 10 | Claude Opus 4 (20250514) | 52.71 | #297 |
| 11 | Grok 3 | 52.18 | #611 |
| 12 | DeepSeek R1 0528 | 51.76 | #428 |
| 13 | Gemini 2.5 Pro (Preview) | 51.02 | #258 |
| 14 | GLM-4.5 | 49.8 | #542 |
| 15 | Qwen 2.5 Coder 7B Instruct | 45.6 | #3914 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=a-s-e-aicgseceval · How It Works · Data refreshed daily, snapshot 2026-11-21.