ExploitGym kernel exploits with mitigations
Agents · 2026-05-11
Count of the 186 ExploitGym Linux-kernel tasks still exploited through the intended vulnerability when the benchmark's kernel defenses are enabled.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-5.6 Sol | 44.0 |
| GPT-5.5 | 8.0 |
| Claude Mythos Preview | 3.0 |
| GPT-5.4 | 1.0 |
| Opus 4.6 | 0.0 |
| Opus 4.7 | 0.0 |
| Gemini 3.1 Pro Preview | 0.0 |
| GLM-5.1 | 0.0 |
| Muse Spark 1.1 | 0.0 |