anthropic.com ↗ Feed Articles Wire
The Life Sciences Verification Program gives approved organizations less restrictive biology safeguards, with monitoring and data retention for flagged activity.
Claude can now route larger tasks, docs, slides, artifacts, connectors and skills from the same conversation instead of making users pick a separate workspace.
Reuters says the pharma company and Anthropic are partnering to use Claude Science and frontier models across research workflows.
The company added finance agent templates, Microsoft 365 add-ins and connectors to market data, research and compliance systems.
Frontier Red Team evaluations found models reaching expert-level performance on some simulated intelligence and conventional weapons workflows.
Anthropic will embed outside evaluators inside the company, while OpenAI's Sam Altman says leading labs may coordinate on safety limits.
The company says it disrupted threat actors using Claude across seven harm areas, from autonomous cyber workflows to biological and conventional-weapons misuse.
The company says a misconfigured evaluation connected models to the open internet; METR will run an independent investigation.
Claude can now draft visual work from a prompt, then hand it off to the user or Claude Code.
Reuters says Nvidia is considering up to 10B USD in the offering, turning Anthropic’s market debut into a test of AI compute economics.
The new Claude model is generally available, while Mythos 5.1 gives vetted cybersecurity and life-science teams less restrictive safeguards.
Anthropic says the model worked largely autonomously for 11 days, turning a famous proof into machine-checkable code.
The company is packaging Claude patterns for consumer and merchant agents as retailers prepare for AI-assisted shopping.
In a controlled research loop, Claude found methods that improved safety benchmarks across 10 failure categories without reducing model capabilities.
Infostealer malware on user devices can copy active sessions, letting attackers drain Claude usage without a normal login.
The publishers say Anthropic illegally copied song lyrics and other works for Claude; Anthropic says it disagrees and will defend the case.
The ruling says the government unlawfully punished Anthropic over its limits on Claude use in weapons and surveillance.
Reuters says the Claude maker discussed buying the AI-chip startup, then shifted toward a possible partnership while building its own silicon team.
The Model Hardware Standard lets AI agents discover and control programmable devices through shared drivers, MCP, command-line tools and APIs.
West Virginia's governor says Bloomberg reported a six-year capacity agreement covering about 460 megawatts at Nscale's planned Monarch Compute Campus.
Anthropic
Nvidia