Claude Opus 5.5 cuts input and output prices 20% and cache reads 60%. Anthropic says typical agentic workloads cost 40% less than with Opus 5.

Claude Opus 5.5 provide Fable 5.1-level performance on most work while Anthropic says typical workloads cost 40% less than Opus 5 with stronger agent safeguards

Anthropic said it blocked scientists who used its Claude AI models "in ways that could support biological weapons development."

Anthropic said Claude now leads 26% of its AI research work, up from under 1% in March, as the industry tracks AI's ability to self-improvement.

This article is based on information as of September 21, 2026.Is testing an AI model only after it is completed sufficient for safety? On September 18, 2026, Anthropic announced that it is collaborating with Accenture to implement 'embedded evaluation,

Anthropic says the firms generated more than 35 million Claude interactions, mostly to improve their AI models, but some Kimi user requests were allegedly routed to the U.S. system without users’ knowledge,

A new Anthropic-focused ETF offers indirect exposure to the AI company through chipmakers, cloud providers and data-center stocks ahead of its expected IPO.

Anthropic's Claude Opus 5.5 system card reveals METR used an undisclosed source in its AI safety evaluation, raising transparency questions in