Hinoki Offensive One: Open Access Cyber Model (hinoki.security)

Hinoki Offensive One (HO1) has emerged as the first open-access cyber model that aims to democratize offensive cyber capabilities for enhanced defensive strategies. Unlike other leading models, such as OpenAI's GPT-5.5-Cyber and Claude Mythos Preview, which refuse to engage in certain tasks, HO1 operates effectively on defined policy parameters, completing a variety of offensive cyber tasks that others do not. This capability makes HO1 a valuable tool for users seeking to test and enhance their cyber defenses. Significantly, Hinoki Offensive One has demonstrated impressive performance metrics, scoring 88.4% during evaluations on CyberGym, surpassing the scores of its competitors by notable margins. This breakthrough not only provides broader access to cybersecurity tools but also encourages innovation and collaboration within the AI/ML community. As it opens new avenues for security professionals and researchers, HO1’s ability to tackle intricate cyber tasks positions it as a pivotal advancement in the landscape of offensive cyber operations and defense.