Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
external benchmark
ExploitGym Hardened (stronger exploit mitigations). Every model the catalog carries a published ExploitGym Hardened value for, ranked by that value.
A hardened variant of ExploitGym that adds address-space randomization and stack canaries to userspace targets, enables the V8 sandbox and uses stricter kernel configurations where supported. 5 models in the ModelScale catalog have a published value.
Each point is a model with a published ExploitGym Hardened value and complete token prices, on the 75% input / 25% output blend. Frontier models are spaced evenly along the cost axis.
5 models with a published ExploitGym Hardened value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Intended-exploit success rate |
|---|---|---|---|
| 1 | OpenAI | 36.9 | |
| 2 | OpenAI | 29.7 | |
| 3 | Anthropic | 5.0 |
What the benchmark measures and how its value is published, from the source’s own record.
ExploitGym Hardened (ExploitGym Hardened (stronger exploit mitigations)) is an external benchmark published by ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks. A hardened variant of ExploitGym that adds address-space randomization and stack canaries to userspace targets, enables the V8 sandbox and uses stricter kernel configurations where supported. Its evaluation set is described by the source as exploitGym tasks with stronger exploit mitigations. The source rates its difficulty as exploitation against mitigations.
Each model's result is published by ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks as intended-exploit success rate. ModelScale prints that value to 1 decimal place, exactly as published, and never converts or rescales it. This page ranks the 5 models that carry a published value, highest value first; a model the source never scored is left out rather than ranked at zero. The source does not state whether a higher value is the better result, so neither does this page.
| 4 | OpenAI | 2.3 |
| 5 | OpenAI | 0.3 |
Evidence key: Observed
Rows are ordered by the value ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks? published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.