Anthropic released Claude Fable 5.1, available immediately on the Anthropic API, AWS, Google Cloud and Azure, alongside a more restricted Mythos 5.1, according to Anthropic's own announcement and TechCrunch.
Fable 5.1 keeps the same $10 per million input token and $50 per million output token pricing as its predecessor, but cuts cached token reads to $0.25 per million, a 75% reduction, Anthropic said. The company estimates 25% lower costs for typical workloads and up to 45% lower for heavily agentic tasks, according to its announcement.
On benchmarks, Fable 5.1 scored 52.6% on Terminal-Bench-Science, more than double the prior model's 24.7%, and 60.9% on Humanity's Last Exam without tools, Anthropic reported. Mythos 5.1 scored higher still on agentic coding but showed a slight regression on overall misaligned behavior compared with Opus 5, the company said.
TechCrunch described Fable 5.1 as the unrestricted version open to all users, while Mythos 5.1 remains limited to registered Anthropic partners doing cybersecurity or life sciences research, currently in the US only. Anthropic reiterated in its post that it "has never trained on enterprise data without explicit permission, and never will."
A 75% cut to cache pricing matters most to anyone running long agentic sessions that repeatedly reread the same context, which is most coding agents and research assistants. That is a direct cost lever for builders already on Claude, not just a benchmark bump.