Anthropic expands Claude Mythos 5 access and launches $35M security fund

Anthropic has announced significant expansions to its Claude AI ecosystem, focusing on cybersecurity and productivity. The company is increasing access to its high-tier cybersecurity model, Claude Mythos 5, through the Claude Security service. To mitigate the risk of the model being used for malicious purposes, Mythos 5 is integrated into specialized interfaces that provide specific defensive outputs, such as vulnerability patches and security alerts, rather than allowing direct, unrestricted prompting.

In tandem with these security updates, Anthropic is launching the Defender Advantage Fund, providing $35 million in Claude credits to support open-source software security and vulnerability detection.

On the productivity front, Claude has introduced enhanced integrations for Google Workspace. The assistant can now manage files in Google Drive and automate email tasks in Gmail, including drafting, replying to, and forwarding messages. While users can opt to disable the ‘Ask for approval’ setting for certain Team and Enterprise accounts, human oversight remains the default for sending communications.