Term
Claude Fable 5.1
Anthropic's flagship Claude Fable model, released on 1 September 2026. A frontier model for demanding reasoning, coding, research and long-running agents; 1M-token context, multimodal input, proprietary access via API and the Claude apps.
Watermarking: this model marks its text
Claude Fable 5.1 is among the models Anthropic ships with provenance marking from 2 August 2026 onwards. Two mechanisms apply: a statistical watermark in the text itself — invisibly woven into word choice, related to Google DeepMind’s SynthID-Text approach — and signed provenance metadata on generated files. Sufficiently long texts therefore carry machine-checkable evidence that Claude was involved.
Important context: the watermark yields a probability, not proof. It needs length, loses signal under heavy rewriting, and code stays largely unmarked — comments, docstrings and documentation do not. Verification runs through a detection interface that is initially available only as a private preview. Older models such as Claude Opus 5 or Sonnet 5 do not mark.
Details: AI Watermarking · Watermarks in AI Text
Claude Fable 5.1 — explained in more detail
Claude Fable 5.1 is the frontier model of Anthropic’s Fable line unveiled on 1 September 2026 and an incremental successor to Claude Fable 5. It is positioned as Anthropic’s most capable generally available model — built for demanding reasoning, long-running agents, coding, research and document work. Alongside Fable 5.1, Anthropic released the closely related Claude Mythos 5.1, a restricted-access variant based on the same underlying weights.
Technically, Fable 5.1 offers a 1-million-token context window with up to 128,000 tokens of output, and accepts images as well as text as input (multimodal). According to the available sources, pricing is around 10 US dollars per million input tokens and 50 US dollars per million output tokens; a headline detail of this release is a roughly 75 percent cut to the cost of cached input, down to about 0.25 US dollars per million tokens.
Access is proprietary: there are no open weights. Fable 5.1 is used through the Claude apps, the Claude API and compatible interfaces.
Example / practical relevance
Fable 5.1 targets the hardest tasks, where a model must work reliably across many steps: analysing and refactoring large code bases, consolidating multi-stage research from many sources, or driving agents that plan autonomously and use tools over hours. The 1-million-token context makes it possible to process very large amounts of documents — an entire content archive or a whole codebase — in a single pass.
The reduced cache cost matters especially for recurring prompts: anyone reusing the same large context (such as a style guide or knowledge document) across many requests pays significantly less for the cached portion.
Distinction from similar terms
Within the Claude generation 5, Fable is the highest-performing frontier model, while Claude Opus 5 comes close to its capabilities but costs less. Claude Mythos 5.1 shares its weights with Fable 5.1 but is access-restricted (invitation/project access). The label 5.1 marks a point update over Fable 5 — not a full generational leap, but improved performance and economics on the same model generation.