What is Claude Fable 5.1?
Claude Fable 5.1 is an Anthropic model for demanding reasoning and long-horizon agentic work, released on September 1, 2026. Anthropic says most workloads should start with Claude Opus 5.5 and move to Fable 5.1 when evaluations on Opus at higher effort still fall short. It costs $10 input and $50 output per million tokens.
What can you do with it?
Anthropic describes Fable 5.1 as extending Claude Fable 5 with improvements in long-running agentic coding, multistep research, and document, spreadsheet and slide work. Its documentation positions it for tasks where Claude Opus 5.5, even at higher effort, does not meet your evaluation bar.
Anthropic labels its latency Slower in the current lineup, so it suits jobs that run unattended rather than interactive chat. Claude Mythos 5.1 offers the same capabilities only to organizations verified through Anthropic's programs.
Fable 5.1 or Opus 5.5?
Anthropic's documentation is direct: for most workloads, start with Claude Opus 5.5. Opus lists $4 input and $20 output per million tokens with Moderate latency and default effort medium, while Fable lists $10 and $50 with Slower latency and default effort high. The context window and output limits are the same at 1M and 128K tokens.
In the Claude apps the gap is about allowance rather than price. Opus is included on Pro and Max, while Fable draws on usage credits on Pro and on half of the weekly limits on Max. Fable 5.1 also adds mid-conversation effort changes in beta and a cache read price a quarter of Fable 5's, which matters for long agent sessions that reuse the same context.
Try it on a small task
Suggested workflow: take a task where Opus 5.5 at high effort already failed your checks, such as a research memo that must reconcile several conflicting sources. Run the same prompt and files on claude-fable-5-1 at its default high effort.
Score both outputs with the same checklist, for example correct citations, no unsupported numbers and every source addressed, and note the token counts. Switch only if Fable passes checks that Opus missed, since each token costs 2.5 times as much.
Pricing and access
On the Claude API, the model ID is claude-fable-5-1. Pricing is $10 per million input tokens and $50 per million output tokens, with cache writes at $12.50 for five minutes or $20 for one hour, cache reads at $0.25 and 50% off on the Batch API. Knowledge cutoff is June 2026, and Anthropic commits to retirement not sooner than September 1, 2027.
In the Claude apps, Anthropic's pricing page lists Fable as unavailable on Free, available through usage credits on Pro and limited to 50% of weekly limits on Max. If you already call Claude Fable 5, four changes are breaking: forced tool use returns an error, only Fable 5.1 and Mythos 5.1 can read its thinking blocks, editing earlier turns invalidates thinking blocks, and some advisor tool pairings are unsupported.
What stands out
- 1M-token context window and 128K max output
- Adaptive thinking always on, default effort high
- Mid-conversation effort changes in beta
- Claude API, Amazon Bedrock, Google Cloud, Microsoft Foundry and Claude Platform on AWS
Before you choose
- Anthropic recommends starting with Opus 5.5 for most workloads.
- Slower latency class than Opus and Sonnet.
- Four breaking changes for code on Fable 5.
Go to the source
Sources & further reading
Based on the official documentation below. Features and access can change; check the provider before making a decision.
Was this page useful?
Help us make our coverage more useful. This is feedback on the page, not a rating of the tool.
Go a little further