Anthropic has announced plans to release its new Mythos-level AI model, which boasts superior intelligence compared to its current Opus model, to all customers in the coming weeks. This development is part of the Project Glasswing initiative, where a select few organizations are already utilizing the Claude Mythos Preview for cybersecurity tasks. However, Anthropic emphasizes the need for enhanced cybersecurity measures before a broader release. Despite the announcement, there is some ambiguity regarding the specific model and the definition of "all customers." On May 23, Anthropic reiterated that while internal tests show the Mythos Preview's exceptional capabilities in cybersecurity, public and commercial access will remain restricted until more robust safety alignment mechanisms are established.