Anthropic Rolls Out Claude Fable 5.1 and Mythos 5.1
Anthropic has pushed out two new additions to its flagship model line, Claude Fable 5.1 and Claude Mythos 5.1, arriving just three months after the original Fable 5 and Mythos 5 debuted. The company is positioning Claude Fable 5.1 as the most capable option on the market for coding and knowledge-heavy work, a claim it made directly on social media platforms alongside the announcement. Both models slot in above the existing Opus tier and are built to handle sustained, multi-step tasks rather than single quick answers.
What Makes Claude Fable 5.1 Different
Claude Fable 5.1 is designed to hold its own across long-running coding projects, deep research assignments, and heavy document review. Anthropic says the model now does a better job tracing problems back to their root cause instead of patching surface-level symptoms, a detail the company backed up with an example involving a financial firm that had struggled for years to pin down an unexplained system crash. The new version reportedly matches or beats its predecessor at lower effort settings while pulling well ahead at the highest effort tiers, and it carries a one-million-token context window with room for responses up to 128,000 tokens.
Lower Costs and New Beta Features
Pricing per token stays flat at ten dollars per million input tokens and fifty dollars per million output tokens, but Anthropic has cut cache-read costs sharply, trimming expenses by roughly a quarter for typical use and by as much as 45 percent for heavily automated workflows. Claude Fable 5.1 also introduces the ability to adjust its effort level mid-conversation rather than locking it in beforehand, along with a content-tracking feature aimed at giving businesses a clearer record of how AI-generated output was produced.
Read more: Fable 5 Export Controls Lifted — A Huge Win for Anthropic
Mythos 5.1 Stays Behind a Gated Program
Claude Mythos 5.1 shares the same underlying architecture as Fable 5.1 but strips away several of its safety layers, which is why Anthropic is keeping it restricted to vetted participants in its Project Glasswing initiative. The company frames this version as suited to specialized cybersecurity and life-sciences research, and it has teased an upcoming Life Sciences Verification Program that would open Mythos 5.1’s biology-related capabilities to a wider pool of scientists once enrollment begins.
A Rocky History With Export Rules
This is not the first time these two models have been in the spotlight. The original Fable 5 and Mythos 5 launched in June, only to be pulled days later after United States export control action tied to cybersecurity concerns. Anthropic restored access once the restrictions were lifted at the end of that month, and the arrival of version 5.1 marks the first major refresh since that disruption.
Timing Lines Up With a Crowded AI Race
The release lands as competition among frontier AI labs keeps accelerating, with rival firms preparing their own next-generation systems. By emphasizing lower running costs, root-cause debugging, and expanded context windows, Anthropic appears to be targeting enterprise teams that lean on AI for extended agentic workflows rather than casual chat use. Whether Claude Fable 5.1 truly holds the advantage it claims will likely come down to how it performs once developers put it through real-world coding and research tasks over the coming weeks.
