Anthropic is dramatically accelerating the availability of Claude Mythos, its latest large language model, by expanding access from a limited initial group to roughly 200 partners globally through Project Glasswing. The expansion covers more than 15 countries, representing a four-fold increase in partner access and signaling that Anthropic believes Mythos has reached sufficient maturity for broader enterprise and developer deployment. This measured rollout appears designed to stress-test the model across diverse use cases and geographies while gathering feedback before a full public launch, a strategy that balances rapid scaling with the company's emphasis on safety validation.
The timing of this expansion coincides with Anthropic's preparation to unveil a new Claude Opus upgrade, suggesting the company is positioning Mythos as a complementary or successor offering within its model lineup. While specific performance benchmarks and capability details remain limited in available reports, the company's confidence in expanding access across regulated markets and diverse partner ecosystems indicates that safety and compliance concerns—traditionally critical gates for Anthropic releases—have been adequately addressed internally. The geographic diversity of the rollout, rather than concentrating in a single market, hints at Anthropic's ambitions to establish Mythos as a global-first offering.
The expansion underscores Anthropic's competitive urgency to demonstrate Claude's ability to scale across enterprise segments and international markets, particularly as competitors refine frontier models. By carefully controlling the partner cohort size and geographic scope during this phase, Anthropic maintains oversight while testing real-world performance and reliability at meaningful scale. This methodical approach reflects the company's founding ethos around AI safety, but also reflects the practical reality that wider adoption requires proven robustness across production environments before mass market release.