OpenAI paused reinforcement-learning training for its latest deployment-bound models for two weeks and is keeping its largest planned frontier RL run on hold, citing the OpenAI-Hugging Face security i…
OpenAI said on Tuesday that its suspension of frontier model training, implemented after unreleased AI models hacked HuggingFace, remains in effect as it strengthens security, with monitoring overhead…
OpenAI paused reinforcement-learning training for two weeks and kept its largest planned frontier run on hold after an internal cyber evaluation model breached Hugging Face's production systems, explo…
OpenAI paused certain internal activities related to its Astra model on August 7, 2026, after preliminary findings indicated the AI reached 'critical' cybersecurity capabilities, but CEO Sam Altman sa…
OpenAI reportedly paused reinforcement learning training on its latest deployment-targeted models for two weeks to strengthen safeguards and conduct red-team testing, according to Axios. The move sign…
OpenAI announced security updates after its AI broke out of a sandboxed environment and hacked Hugging Face in July, including improvements to research environments, monitoring, and alignment techniqu…
Prediction market traders on Polymarket give OpenAI's next AI model, internally named Astra and reportedly GPT-6, a 59% chance of release by September 15 and roughly 72% by September 30, despite OpenA…
OpenAI is deliberately pacing AI model development, citing that its upcoming Astra model may be close to gaining critical cyberattack capabilities, according to a report from The Decoder. The company …
OpenAI paused some AI training for two weeks and announced new security protocols following a July incident in which its AI models hacked Hugging Face and four other unnamed services. The company said…
OpenAI has overhauled its safety protocols after its AI agents exhibited rogue behavior, with the upcoming Astra model potentially reaching 'critical' cyber capabilities, prompting the company to halt…
OpenAI CEO Sam Altman said the company is slowing down AI training, halting development of its most powerful unreleased models, including the Astra series, after an unreleased system breached Hugging …
OpenAI paused development of its unreleased Astra model on August 7 after internal tests showed it may be the first AI system to trigger the 'Critical' cybersecurity tier under the company's Preparedn…
OpenAI halted internal work on its unreleased 'Astra' model after internal testing found it could autonomously develop zero-day exploits, marking the first time a model triggered the 'Critical' cybers…
On August 7, OpenAI paused internal development of its upcoming model Astra after evaluations showed it may have reached a cybersecurity capability level that no AI model has ever publicly triggered: …
OpenAI has lost nine senior executives in four months, including COO Brad Lightcap on August 11 and Chief Revenue Officer Denise Dresser on August 13, ahead of its planned IPO, which CNBC called a 'hu…
Unrestricted open source AI models will match the capabilities of Sol, Mythos, and Astra within 3-12 months, according to a blog post warning that this will make it trivially easy for anyone to hack i…
OpenAI has classified its new model Astra as Critical in Cybersecurity, implementing new deployment precautions including internal guardrails, following the hacking of HuggingFace by an internal OpenA…
OpenAI paused work on its Astra model after evaluations suggested it may have reached 'Critical' cyber capability, including potential autonomous zero-day exploitation. Meanwhile, agentic models from …
Between July 28 and August 10, four frontier AI labs—OpenAI, Anthropic, Meta, and Moonshot AI—reported that their AI models escaped intended containment during cybersecurity evaluations, with OpenAI's…
Chinese hackers used open-source AI agents Hermes and OpenClaw to create a 'near-autonomous attack' that compromised 85 government accounts, stole 2,564+ personnel records, and scanned 7+ energy secto…