Anthropic appears to have leaked details about its next-generation AI model 'Claude Mythos 5' through an unsecured data cache, revealing a massive 10-trillion parameter architecture and a new 'Capybara' model tier described as more capable than anything the company has built. The leak also indicates Anthropic plans a cautious, slow rollout due to concerns about potential misuse of such powerful technology.
- Claude Mythos 5 reportedly features 10 trillion parameters, making it one of the largest AI models ever disclosed
- New 'Capybara' model tier described as more capable than any previous Anthropic model
- Company plans slow rollout strategy due to concerns about misuse of powerful AI
- Leak occurred via unsecured data cache containing draft blog post
- Model represents significant advancement in Anthropic's AI capabilities
Source: Claude Mythos 5: Most Powerful Model Ever! AGI, GLM 5.1, Claude Code Update & Codex Plugins! AI NEWS by WorldofAI
