Anthropic has officially unveiled Claude Fable 5.1, and this is not just an incremental update. Alongside the new model, the company has introduced a revolutionary protection system that will forever change the rules of the game in the industry: now, new developer accounts are stripped of the ability to rewrite conversation history, while keeping the original "thinking" of the Claude model unchanged.
This move is a direct response to the widespread problem of distillation, where competitors use the API of expensive models to train their own cheaper alternatives. The company emphasizes that Fable 5.1 and the limited Mythos 5.1 are the most advanced solutions for programming and knowledge-related tasks. Both products are already available on key cloud platforms, including Amazon Bedrock, Google Cloud, and Microsoft Foundry.
Price and Performance: What's New
Base rates remain unchanged: $10 per million input tokens and $50 per million output tokens. However, the economic benefit now lies in the details. Cache reads have dropped by 75% — to $0.25 per million operations. According to Anthropic's estimates, typical tasks cost 25% less, and complex scenarios for AI agents save up to 45%.
The performance breakthrough is evident. In the Terminal-Bench-Science 0.1 test, the model scored 52.6%, while the previous version, Fable 5, barely reached 24.7%. In programming, the result is also impressive: 55.8% in Terminal-Bench 4.0, significantly ahead of competitors, including OpenAI's GPT-5.6 Sol with 37.3%. The model's knowledge base has been expanded to June 2026 — five months "fresher" than its predecessor.
"The price is the same, but cache queries are now 75% cheaper. The model works longer on a task before seeking help and better signals dead-end situations," the official developer channel notes.
Protection Against Distillation: A Strategic Maneuver
The problem of copying has long been brewing. As early as February, Anthropic tracked over 16 million conversations related to distillation through approximately 24,000 fake accounts. The company directly named three Chinese labs — DeepSeek, Moonshot AI, and MiniMax — which accounted for millions of requests. In July, White House Director Michael Kratsios publicly accused Moonshot of copying the flagship model to create Kimi K3.
The new restriction applies only to accounts created after August 31. Existing users of Claude Code, Cowork, and Claude.ai will not notice any changes. However, Anthropic warns that in future versions, all accounts will undergo verification. Fable 5 will remain available until at least June 2027, so there is no need to rush migration.
My analysis: This is not just a technical update but a clear signal to the market. Anthropic is turning intellectual property protection into a competitive advantage, forcing opponents to seek new paths. In the long term, such measures could slow the pace of AI development in countries relying on distillation, but for the industry as a whole, it is a step toward a fairer and more sustainable ecosystem.