Summary
On September 1, 2026, Anthropic released Claude Fable 5.1, its improved general-purpose agent model, alongside a restricted Mythos 5.1 for vetted cyberdefense and life-sciences organizations. Fable 5.1 keeps Fable 5 base rates but cuts cache-read pricing 75%, from $1 to $0.25 per million tokens, and defaults to a 1M-token context window.
What changed
Fable 5.1 (API id claude-fable-5-1) ships with a 1M-token context window as both default and max, 128k max output, unchanged base rates of $10/M input and $50/M output, and a cache-read price cut from $1 to $0.25 per million. Anthropic estimates roughly 25% savings on typical workloads and up to 45% on heavily agentic ones. Mythos 5.1 is gated to vetted defenders and researchers via trusted-access programs.
Why it matters
The cache-read cut and flat 1M-context pricing target the exact cost driver of long-running, stateful agent sessions, making multi-hour agent runs and large-context code and research workflows meaningfully cheaper. It is the model already flowing into GitHub Copilot and Vercel AI Gateway, so the price change ripples across the tools developers already use.
Evidence excerpt
Fable 5.1 keeps Fable 5 base rates ($10/M input, $50/M output) but cuts cache reads 75% from $1 to $0.25 per million; a 1M-token context window is both default and max with flat per-token pricing and 128k max output. Mythos 5.1 is restricted to vetted cyberdefense and life-sciences organizations.