01
Release and availability
Anthropic documents Claude Fable 5.1 as released on September 1, 2026 and active as its latest Fable model. The Claude API identifier is claude-fable-5-1.
Anthropic lists availability through the Claude API, Amazon Bedrock, Google Cloud, Microsoft Foundry and Claude Platform on AWS. Claude Mythos 5.1 shares the same underlying capabilities but is invitation-only through Project Glasswing; it should not be treated as generally available.
02
Context window, output and reasoning
The current model documentation lists a 1 million-token context window and up to 128K output tokens. Adaptive thinking is always on, with effort controlling reasoning depth.
A large context window is useful for codebases, research corpora and long agent histories, but teams should still evaluate retrieval and task accuracy on their own data.
03
Claude Fable 5.1 API pricing
Anthropic currently lists $10 per million input tokens and $50 per million output tokens. Five-minute cache writes cost $12.50/M, one-hour cache writes cost $20/M, and cache reads cost $0.25/M. Batch API input and output receive a 50% discount.
Base input and output rates are unchanged from Fable 5. The important pricing change is cache reads at one quarter of Fable 5’s rate, which can matter for long-running agents that repeatedly reuse large cached prefixes.
04
What changed from Fable 5
Anthropic documents three breaking changes for existing Fable 5 callers: forced tool use can return an error, earlier models cannot read Fable 5.1 thinking blocks, and editing earlier turns invalidates thinking blocks.
Additions include per-message effort, turn-scoped system messages, readable progress updates between tool calls, lower cache-read pricing and content provenance. Production agents should follow Anthropic’s migration guide instead of blindly swapping the model ID.
05
How to read Anthropic’s benchmark claims
Anthropic’s launch materials report gains over Fable 5 across coding, research and automation evaluations. These are vendor-published benchmark results, not independent proof of universal superiority.
Harnesses, effort settings, tools, safeguards and task mixes can materially change results. For production decisions, compare task success, latency and total cost on a representative internal evaluation set.
06
Where Fable 5.1 fits
Anthropic’s model guide says most workloads should start with Claude Opus 5 and positions Fable 5.1 for demanding reasoning and long-horizon agentic work when Opus 5 at higher effort still falls short.
That makes Fable 5.1 most relevant to difficult coding agents, multistep research, large-document work and workflows where cached context is repeatedly reused. Choose it on task-level quality and total cost per successful outcome rather than model tier alone.
07
Data retention and access boundaries
Anthropic’s current documentation says Fable 5.1 requires 30-day data retention and is not available with zero data retention unless Anthropic expressly authorizes it. Review the current provider and organizational terms before sending sensitive production data.
Sources
Primary and supporting sources
Facts were rechecked against the linked sources immediately before publication. Pricing, product availability and rollout status can change.