Daily updates on AI news and trends worldwide
Anthropic is a US AI company founded in 2021 with a focus on AI safety. It develops the Claude family of large language models and offers conversational AI and the Claude Code coding tool through its API and apps.
| Model | Input ($/1M tokens) | Output ($/1M tokens) | Released | Source |
|---|---|---|---|---|
| Claude Fable 5 | $10.00 | $50.00 | 2026-06-09 | Source |
| Claude Mythos 5 | $10.00 | $50.00 | 2026-06-09 | Source |
| Claude Opus 5 | $5.00 | $25.00 | 2026-07-24 | Source |
| Claude Opus 4.8 | $5.00 | $25.00 | — | Source |
| Claude Opus 4.7 | $5.00 | $25.00 | — | Source |
| Claude Sonnet 5 | $2.00 | $10.00 | 2026-06-30 | Source |
| Claude Haiku 4.5 | $1.00 | $5.00 | — | Source |
Anthropic's Frontier Red Team ran multiple Claude agents in shared environments. Coordinated swarms found 266 vulnerabilities vs 21 for independent agents, but conformity, collusion, and sabotage also emerged.
Anthropic signed the EU AI Act Article 50(2) transparency code and will weave an imperceptible watermark into text Claude generates, plus C2PA-signed provenance metadata on generated files. Models launched in the EU on or after August 2, 2026 are covered, and marking applies worldwide.
Anthropic released self-hosted environments for Claude Code in public beta on August 6, 2026, letting sessions run inside an organization's own network. Available to Claude Team and Enterprise organizations, off by default.
Anthropic released an inline DLP feature for Claude Enterprise on August 5, 2026. Prompts and tool responses wait for a verdict from the organization's own security server before reaching the model. Available in beta.
Anthropic will make auto mode the default for Claude Code on Pro, Max, and Team plans starting August 14. The company published a controlled study of 1,053 testers comparing human approval with its classifier. Enterprise stays opt-in for now.
Claude Code v2.1.224, published to npm on August 7, 2026, lets sessions send each other messages. Only plain text crosses over — never conversation history or files — and the feature runs on macOS and Linux.
Anthropic said on August 7, 2026 that it rewrote and retrained the classifier governing Fable 5's biology safeguards. Everyday health and educational questions trigger fewer fallbacks, while virology, toxicology and molecular design still route to Opus 5.
Anthropic has posted roles for a custom silicon team. The Silicon Engineer role pays $320,000-$485,000 and asks for engineers who have shipped silicon; a chip-design RL role pays $500,000-$850,000. The company says it will keep relying on AWS, Google, Nvidia and AMD hardware.
Anthropic announced on August 4, 2026 that Mariano-Florentino (Tino) Cuéllar will join as its first Chief Global Affairs Officer, leading policy, international engagement, and government relations.
A White House official confirmed on August 3, 2026 that a voluntary framework for reviewing frontier models' cybersecurity capabilities is complete, with a meeting with leading AI companies set for the next day. The framework and its benchmarks stay classified.
After July's containment breakouts at OpenAI and Anthropic, AFP asked legal scholars who is responsible. A look at what the statutes actually say: the CFAA's intent requirements, and the single defense California's AB 316 removed.
An open letter signed by Anthropic CEO Dario Amodei, SSI's Ilya Sutskever and other frontier AI company employees asks the US government to support an international effort to build the technical and governance tools needed to deliberately pace automated AI development.