SecCodeBench: leaderboard
Security benchmark for AI-generated and AI-repaired code, reporting secure-code repair and generation scores with and without hints.
Metric: Total Score. Source: alibaba.github.io. Saturation forecast: Around February 2027. 36 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.5 | 68.1 | #83 |
| 2 | Qwen 3.5 Plus (Thinking) | 67.08 | #125 (Qwen 3.5 Plus) |
| 3 | Qwen 3.7 Max (Thinking) | 66.81 | #75 (Qwen 3.7 Max) |
| 4 | Claude Opus 4.6 | 64.9 | #63 |
| 5 | Qwen3 Coder Next | 62.76 | #326 |
| 6 | Gemini 3 Pro | 62.42 | #80 |
| 7 | GLM-5 (Thinking) | 62.13 | #141 (GLM-5) |
| 8 | Kimi K2.5 (Thinking) | 61.25 | #144 (Kimi K2.5) |
| 9 | GPT-5.4 | 59.74 | #79 |
| 10 | Qwen 3.6 Max (Preview) (Thinking) | 58.7 | |
| 11 | Gemini 3 Flash | 58.66 | #99 |
| 12 | GPT-5.2 | 58.23 | #110 |
| 13 | Claude Sonnet 4.5 | 56.83 | #142 |
| 14 | DeepSeek V3.2 (Thinking) | 55.24 | #205 (DeepSeek V3.2) |
| 15 | Kimi K2.5 (Non-reasoning) | 55.22 | #144 (Kimi K2.5) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=seccodebench · How It Works · Data refreshed daily, snapshot 2026-10-20.