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
Every model the catalog carries a published FrontierCyber value for, ranked by that value.
Independent evaluation of AI agents against vulnerable real-world systems in dynamic environments. 3 models in the ModelChoice catalog have a published value.
Each point is a model with a published FrontierCyber value and complete token prices, on the 75% input / 25% output blend. Frontier models are spaced evenly along the cost axis.
3 models with a published FrontierCyber 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 | Tasks solved |
|---|---|---|---|
| 1 | OpenAI | 38.1 | |
| 2 | OpenAI | 31.3 | |
| 3 | OpenAI | 9.6 |
Evidence key: Observed
What the benchmark measures and how its value is published, from the source’s own record.
FrontierCyber is an external benchmark published by FrontierCyber. Independent evaluation of AI agents against vulnerable real-world systems in dynamic environments. Its evaluation set is described by the source as 197 dynamic cyber tasks. The source rates its difficulty as easy through elite cyber operations.
Each model's result is published by FrontierCyber as tasks solved. ModelChoice prints that value to 1 decimal place, exactly as published, and never converts or rescales it. This page ranks the 3 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.
Rows are ordered by the value FrontierCyber 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.