Anthropic announced it has suspended user access to its two newest language models, Claude Mythos 5 and Claude Fable 5. The decision appears on the company’s status page and was quickly picked up by the developer community on Hacker News. Both models were positioned as upgrades to earlier Claude releases, promising larger context windows and refined instruction following.

The outage affects developers who integrated the models via the public API, forcing them to fall back on older versions or pause production workloads. Users reported error messages when attempting calls, and the status page lists the incident as ongoing without a projected resolution time. The rollback highlights the fragility of relying on cutting‑edge AI services for mission‑critical pipelines.

With the suspension, teams must audit code paths that reference the affected endpoints and consider alternative providers if uptime is essential. Anthropic has not detailed the cause, but the incident underscores the need for robust fallback strategies when adopting rapidly evolving generative AI tools.