Atlas

Benchmarks

← All benchmarks

ExploitGym V8 exploits with mitigations

Agents · 2026-05-11

Count of the 181 ExploitGym V8 tasks still exploited through the intended vulnerability when the V8 heap sandbox and other benchmark defenses are enabled.

Top models (higher is better)

ModelScore
GPT-5.6 Sol32.0
Claude Mythos Preview17.0
GPT-5.54.0
Opus 4.60.0
Opus 4.70.0
Gemini 3.1 Pro Preview0.0
GLM-5.10.0
GPT-5.40.0
Muse Spark 1.10.0
Loading Atlas data…