A.S.E (AICGSecEval) - Security: leaderboard
Code security score (%) on Tencent's A.S.E (AICGSecEval, arXiv:2508.18106v3, Tables 2-3), measuring the share of repository-level CVE code completion tasks where generated code reduces tailored static-analysis vulnerability alerts.
Metric: Code security score (%). Source: arxiv.org. Saturation forecast: Not forecast. 33 models tracked. Updated 21 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 235B A22B 2507 Instruct | 48.03 | #616 |
| 2 | Claude 3.7 Sonnet (20250219) | 46.72 | #421 |
| 3 | GPT-4o (2024-11-20) | 45.65 | #859 |
| 4 | Qwen 3 30B A3B 2507 Instruct | 45.46 | #1100 |
| 5 | Qwen 3 Coder 480B A35B Instruct | 42.69 | #652 |
| 6 | DeepSeek V3 (0324) | 40.89 | #715 |
| 7 | Qwen 2.5 Coder 7B Instruct | 40.78 | #3914 |
| 8 | Grok 3 | 38.64 | #611 |
| 9 | DeepSeek R1 0528 | 38.01 | #428 |
| 10 | Qwen-Coder-Plus-20241106 | 37.98 | |
| 11 | Kimi K2 (Preview) | 37.82 | #307 |
| 12 | GLM-4.5 | 35.92 | #542 |
| 13 | Claude Sonnet 4 (20250514) | 34.78 | #405 |
| 14 | Qwen 3 4B 2507 (Thinking) | 33.57 | #1433 |
| 15 | Qwen 2.5 Coder 14B Instruct | 32.24 | #1832 |
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-security · How It Works · Data refreshed daily, snapshot 2026-11-21.