Model trained on 150,000 CVE exploits scores 88.4% on CyberGym with zero refusals
jedisct1 · x · 2026-11-09
Johnny Wang announces Hinoki Offensive One, an LLM trained on 150,000 hand-written CVE exploits. It scores 88.4% on CyberGym — 2.8 points above Daybreak — with zero refusals, and is pitched as unrestricted pentesting and exploit development, available now.
More from Models
- Grok API text outputs now watermarked by default across 7 models for EU AI Act — thione · 2026-11-10
- Moonshot AI open-sources Kimi K3.1 Pro: 2.8T-param multimodal agent model with 1M-token context — thione · 2026-11-10
- OpenAI releases GPT-6 Cyber to approved security teams, expanding exploit-development capabilities — thione · 2026-11-10
- RT claim: GPT-4-level intelligence has gotten roughly 1800x smaller — CatAstro_Piyush · 2026-11-10
- Private OCR eval on 338 scanned government pages: 0.8B model beats 3B-class rivals — vanstriendaniel · 2026-11-10
- Kimi and 1017 are closer in performance than they appear, and both trail the frontier — teortaxesTex · 2026-11-10