Summary
On September 1, 2026, Anthropic released Claude Fable 5.1, with Mythos 5.1 as the higher-safeguard variant of the same model, tuned for coding, knowledge work, and long-running agentic tasks. Headline token pricing stays at $10 per million input and $50 per million output, but cache-read pricing drops 75% to $0.25 per million tokens, cutting typical workload cost roughly 25% and highly agentic work up to about 45%.
What changed
Anthropic shipped Fable 5.1 and Mythos 5.1, replacing Fable 5, with higher benchmark scores such as 52.6% on Terminal-Bench-Science versus 24.7% for Fable 5, and a 75% cut to cache-read pricing to $0.25 per million tokens.
Why it matters
The cache-read price cut targets the dominant cost driver in agentic and long-context workloads, where repeated context reads accumulate across a run. Cheaper caching plus stronger long-horizon performance makes frontier-tier models more economical for coding agents and multi-step automation, shifting competition toward cost-per-agentic-task rather than headline token rates.
Evidence excerpt
Fable 5.1 keeps the same headline rate as Fable 5, at $10 per million input tokens and $50 per million output tokens. However, cache-read pricing drops from $1.00 to $0.25 per million tokens, a 75% reduction.