Fable 5 is back and available (platform.claude.com)

Anthropic has restored access to its advanced AI models, Claude Fable 5 and Claude Mythos 5. Claude Fable 5 is now available as the most capable model for intricate reasoning and long-horizon tasks, while Claude Mythos 5, which shares similar capabilities, is released under strict conditions through Project Glasswing. Notably, Claude Fable 5 features safety classifiers designed to decline certain requests, a significant enhancement for responsible AI usage, whereas Claude Mythos 5 does not include these classifiers. This update is crucial for developers and companies integrating AI into their platforms, as it introduces new operational protocols. When Claude Fable 5 declines a request, it returns a "refusal" response, necessitating integration changes that include handling these refusals, establishing fallback mechanisms to retry with other Claude models, and adhering to adjusted billing rules—important for cost management. Both models maintain consistent specifications and pricing, but the safety features of Claude Fable 5 highlight a progressive step towards ensuring that AI applications behave more ethically and responsibly. This launch, alongside detailed instructions for migration and utilization, positions Claude Fable 5 at the forefront of current AI capabilities.