The AI firm said that unlike previous model bugs, this issue "crept in subtly".

If OpenAI can accidentally train its flagship model to obsess over goblins, what other more subtle and potentially harmful ...

FRANKFURT, March 31 (Reuters) - The European Central Bank has developed models to more closely monitor shifts in inflation expectations, it said in a blog post on Tuesday, a potentially crucial ...

In a statement provided to Fortune after the leak, an Anthropic spokesperson described Claude Mythos as an AI performance “step change” and said the model is “the most capable we’ve built to date.” ...

OpenAI on Thursday released its latest flagship model, GPT-5.5, according to a blog post. That model, codenamed Spud internally , is better at breaking down complicated requests from customers and ...

The draft blog post describes a compute‑intensive LLM with advanced reasoning that Anthropic plans to roll out cautiously, starting with enterprise security teams. Anthropic didn’t intend to introduce ...

OpenAI says its safeguards “sufficiently reduce cyber risk” for now, while GPT-5.4-Cyber is a new cybersecurity-focused model.