CyberGym-E2E Patch-only (920 tasks)
Agents · 2026-06-03
CyberGym-E2E patch-only success rate on the current 920-task set. Agents receive the ground-truth proof of concept and crash log, isolating their ability to diagnose and patch the vulnerability.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-5.4 | 87.1 |
| Opus 4.6 | 85.8 |
| Gemini 3.1 Pro Preview | 83.0 |