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
Atomic Vulnerability Research and Exploitation. Every model the catalog carries a published Atomic vulnerability research value for, ranked by that value.
Irregular's domain-level evaluation of vulnerability research and exploitation capability. 3 models in the ModelScale catalog have a published value.
Each point is a model with a published Atomic vulnerability research 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 Atomic vulnerability research 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 | Domain average |
|---|---|---|---|
| 1 | OpenAI | 100 | |
| 2 | OpenAI | 92 | |
| 3 | OpenAI | 91 |
What the benchmark measures and how its value is published, from the source’s own record.
Atomic vulnerability research (Atomic Vulnerability Research and Exploitation) is an external benchmark published by Assessing GPT-5.6 Sol. Irregular's domain-level evaluation of vulnerability research and exploitation capability. Its evaluation set is described by the source as atomic cyber tasks. The source rates its difficulty as vulnerability research and exploitation.
Each model's result is published by Assessing GPT-5.6 Sol as domain average. ModelScale prints that value to 0 decimal places, 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.
Evidence key: Observed
Rows are ordered by the value Assessing GPT-5.6 Sol 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.