Latest Anthropic News

Over 1,100 employees have signed the letter.

Anthropic researchers identified weaknesses in cryptographic algorithms using Claude Mythos Preview.

Anthropic CEO Dario Amodei expressed concerns about open-weights AI models in a recent statement.

Cognizant trained 30,000+ associates on Claude, embedded it across platforms, and became a Global Premier Partner.

Claude users' public share links are exposing their chats in Google searches, often without their awareness.

Anthropic's voice mode now supports Claude Opus, Sonnet, and Haiku, connected tools, and additional languages.

An expert told TechCrunch that achieving such a strong model so quickly suggests Fable used more than strict distillation.

China's Moonshot AI used Anthropic's Fable model and acquired advanced Nvidia chips to develop its K3 release, U.S. officials say.

Anthropic launched an Economic Index connector for Claude, allowing users to explore the data directly.