The New Stack

Latest developments

2026

◆ OpenAI released GPT-6 Astra after the model met the “Critical” cybersecurity capability threshold in its Preparedness Framework Model release

◆ Google Research releases TimesFM-3, a zero-shot multivariate time-series forecasting foundation model Model release

◆ Anthropic adds the Claude Mythos 5 model to its Claude Security enterprise vulnerability scanning tool in public beta Feature update

◆ OpenAI reorganized its internal preparedness function and paused some reinforcement-learning training to tighten security and safeguards Security issue

• 2 Sep Your next OpenAI API timeout might not be a timeout at all
• 31 Aug Google’s new forecasting model beats everyone. You can’t use it at work (yet).
• 27 Aug “Posterity will find it ludicrous”: Sai agent hits 73% on OSWorld 2.0 performing routine (but necessary) work
• 26 Aug Observability has a data problem. AI is about to make it worse.
• 21 Aug Anthropic brings Mythos 5 to its Claude Security vulnerability scanner
• 19 Aug “The opening stages of OpenAI’s unraveling”: OpenAI slows model training — not everyone is buying the explanation

Relationships

Partnerships

• OpenAI partnered with this industry · 1 source