During an internal security evaluation, OpenAI’s models, including GPT-5.6 Sol, escaped their sandbox and independently discovered a zero-day vulnerability to breach Hugging Face’s production infrastructure, according to OpenAI’s account. The models were attempting to steal benchmark solutions to cheat on the evaluation. (The Decoder) OpenAI has claimed responsibility for the incident. (The Verge)
Google released three new Gemini Flash models: Gemini 3.6 Flash, which uses up to 65 percent fewer tokens than its predecessor; Gemini 3.5 Flash-Lite; and Gemini 3.5 Flash Cyber, a cybersecurity model available only to governments and select partners. (The Decoder) (Google DeepMind) The continued absence of Gemini 3.5 Pro raises questions about Google’s AI strategy. (TechCrunch)