BLXBench Security category
Code · 2026-04-25
Security category component of the BLXBench public leaderboard. The source describes this category as secure code changes, vulnerability recognition, and safe defaults; scores are category score percentages.
Top models (higher is better)
| Model | Score |
|---|---|
| Opus 4.7 | 83.6 |
| Opus 4.8 | 80.7 |
| MiniMax M3 | 75.4 |
| GPT-5.5 | 74.8 |
| GPT-5.3-Codex | 73.0 |
| Step 3.7 Flash | 73.0 |
| Kimi K2.7 Code | 71.1 |
| Kimi K2.6 | 70.1 |
| MiMo-V2.5-Pro | 68.0 |
| GLM-5.1 | 65.6 |
| Qwen3.7-Max | 64.5 |
| GLM-5.2 | 64.1 |
| Nemotron 3 Super 120B A12B | 64.1 |
| Grok Build 0.1 | 63.1 |
| CoBuddy | 61.7 |
| Gemini 3.1 Flash-Lite | 61.3 |
| MiniMax M2.7 | 60.7 |
| DeepSeek-V4-Pro | 60.4 |
| Mistral Medium 3.5 | 60.2 |
| DeepSeek-V4-Flash | 58.2 |
| Nemotron 3 Nano 30B A3B | 57.2 |
| MiMo-V2.5 | 56.3 |
| Qwen3.7-Plus | 55.3 |
| Ring 2.6 1T | 54.7 |
| North Mini Code | 54.5 |
| Qwen3.6-Flash-2026-04-16 | 52.3 |
| Grok 4.3 | 48.4 |
| Gemini 3.5 Flash | 46.1 |
| Mistral Small 4 | 46.1 |
| Nemotron 3 Nano Omni 30B A3B | 44.9 |
| Granite 4.1 8B | 44.1 |
| Claude Fable 5 | 40.6 |