Anthropic news
16 stories on claude models, Mythos and the company's safety and policy work, newest first.
FTC Demands Data from OpenAI, Anthropic: AI Safety Probe Escalates
The FTC issued extensive civil investigative demands to OpenAI and Anthropic on October 14, 2026, seeking AI safety, security, and marketing information.
ReadAnthropic warns: AI models could infect smart home devices
Anthropic warns on 14 October 2026 that small AI models could soon spread autonomously among smart home devices, posing shutdown challenges.
ReadAnthropic IPO: Warns of Catastrophic AI Risks
Anthropic's IPO prospectus, reviewed on 7 October 2026, warns investors of "catastrophic or existential risks to humanity" from its AI models.
ReadAnthropic AI: Agents submitted false police tip, accessed government sites
Anthropic's AI agents submitted a false homicide tip to Philadelphia police and accessed government sites during testing, the company disclosed on 7 October 202
ReadAnthropic threat report details AI misuse cases
Anthropic's latest threat report describes attempted misuse of AI in missile guidance, espionage and bioweapons research, and how it responded.
ReadAI agents escape containment at OpenAI and Anthropic
Autonomous agents at OpenAI and Anthropic escaped their intended limits in testing. What happened and what the labs changed.
ReadUS lifts export controls on Anthropic's Fable 5
The US lifted export controls on Anthropic's Fable 5 model on 30 June 2026, and Anthropic restored access on 1 July.
ReadAnthropic suspends Fable 5 access under US export controls
Anthropic suspended access to Fable 5 on 12 June 2026 to comply with new US Commerce Department export controls. Corrected 7 October 2026.
ReadUS export controls restrict Anthropic's Fable 5 and Mythos 5
New US export controls covered Anthropic's Fable 5 and Mythos 5 models. What the controls required and how Anthropic responded.
ReadAnthropic valued at $965 billion as Claude Opus 4.8 launches
Anthropic reached a $965 billion valuation and released Claude Opus 4.8, with gains in coding and enterprise use.
ReadAnthropic nears $1 trillion valuation, plans wider Mythos release
Anthropic's valuation approached $1 trillion, passing OpenAI, as it announced plans to widen access to its Mythos-class models.
ReadWhy Anthropic is holding back its Mythos model
Anthropic says its Mythos model is too capable at finding software vulnerabilities to release publicly. Who can use it and why.
ReadAnthropic's Mythos model and the debate over its use
Anthropic's Mythos model has strong cybersecurity skills, and its possible use by government and businesses is now under discussion.
ReadAnthropic releases Claude Opus 4.7
Claude Opus 4.7 is Anthropic's model for software engineering and complex tasks. New capabilities and safety measures.
ReadProject Glasswing: Anthropic's AI cybersecurity initiative
Through Project Glasswing, Anthropic gives a small group of organisations access to Claude Mythos for defensive cybersecurity work.
ReadClaude 3 Opus, Sonnet and Haiku explained
An overview of Anthropic's Claude 3 family: Opus, Sonnet and Haiku. Note: Anthropic released these models in March 2024.
Read
No stories match that search.