A US model-evaluation and threat-research organization said on the 5th that Anthropic's AI models have recently improved task-specific performance at a speed that may be about twice prior levels. Anthropic CEO Dario Amodei publicly urged in September slowing the development of frontier AI to allow safety measures to keep pace.

2026-11-06

A US model-evaluation and threat-research organization said on the 5th that Anthropic's AI models have recently improved task-specific performance at a speed that may be about twice prior levels. Anthropic CEO Dario Amodei publicly urged in September slowing the development of frontier AI to allow safety measures to keep pace.