🟢 Claude Fable 5 · model notice · Fable subscription
Fable 5 and 5.1 subscription (updated September 7, 2026): Claude Fable 5.1, released September 1, 2026, is the current Fable model and Fable 5 is now legacy. Plan terms are the same for both — Max and Team Premium plans include Fable at up to 50% of the weekly usage limit; Pro and Team Standard use usage credits (
Update (September 27, 2026): Claude Opus 5.5 was released on September 22, 2026 and is now the current Opus model (API claude-opus-5-5, $4 per million input tokens and $20 per million output tokens, down from Opus 5's $5/$25). Anthropic says it performs at the level of Fable 5.1 on most work and costs 40% less to run than Opus 5 on typical workloads. Opus 5 remains available but is marked legacy in the official model docs. Opus 5 prices and comparisons below reflect the pre-launch state and will be updated. Details: Claude Opus 5.5 complete guide.
Anthropic released Claude Fable 5.1 and Claude Mythos 5.1 on September 1, 2026. This guide focuses on one question — "what does a Fable 5 user need to know?" — and covers the changed specs, pricing, subscription terms, and API migration caveats, all based on the official docs. It takes about 6–7 minutes to read.
What Fable 5.1 is
The official model docs describe Fable 5.1 as an extension of Fable 5 at the same input and output prices. Three areas are strengthened:
- Long-running agentic coding — development work where the model plans multiple steps and calls tools over an extended session
- Multistep research — investigations that loop through searching, organizing, and verifying
- Document, spreadsheet, and slide work — the office deliverables that dominate knowledge work
The official recommendation is also clear: start with Opus 5 for most workloads, and use Fable 5.1 for demanding reasoning and long-horizon agentic work, or when your evals on Opus 5 at higher effort (the setting that controls how deeply the model thinks) still fall short. In other words, Fable 5.1 is the top tier you step up to, not the default.
Mythos 5.1 shares Fable 5.1's specifications and pricing but is offered by invitation only to Project Glasswing participants (a cyber-defense research program). The model open to general users and developers is Fable 5.1.
Specs and pricing — what's the same as Fable 5, what's different
Reading the chart above line by line:
- Unchanged — $10 per million input tokens, $50 per million output tokens. A 1M-token context window (how much it handles at once) and 128K max output. Adaptive thinking (the model decides on its own when to think deeply) is the only thinking mode. The Batch API 50% discount also carries over.
- Cheaper — the cache read price (prompt caching — a discount when you resend the same prefix) dropped from $1 to $0.25 per million tokens. That is 0.025x the base input price, versus the 0.1x multiplier on every other model. The longer the system prompt or document you resend on every turn, the larger the saving. Cache write prices are unchanged: $12.50 for 5-minute and $20 for 1-hour caches.
- Newer — the reliable knowledge cutoff (the point up to which the model's knowledge is dependable) moved from January 2026 to June 2026. Default effort is high.
- Support window — the docs list Fable 5.1's retirement as not sooner than September 1, 2027. Fable 5 is now a legacy model: still available, but the docs recommend migrating to 5.1.
Platforms: the Claude API, Amazon Bedrock (model ID anthropic.claude-fable-5-1), Google Cloud, Microsoft Foundry, and Claude Platform on AWS.
How it works on subscription plans
The official help article states that Fable 5 and Fable 5.1 work the same way on your plan. By plan:
- Max, Team premium seats, Enterprise premium seats — included as a standard part of the plan. You can spend up to 50% of your weekly usage limit on Fable models; beyond that, continue with usage credits (pay-as-you-go beyond what the plan includes) or switch to another model. This is not a 50% larger limit — it is a 50% share of the same weekly limit.
- Pro, Team standard seats, Enterprise standard seats — not included in plan limits; Fable models run on usage credits from the start.
- Free — not available.
One difference matters: the one-time $100 credit that eligible Pro and Team standard users received when Fable 5 moved to usage credits in July 2026 does not apply to Fable 5.1. The help article states that Fable 5.1 was never part of that promotion.
To use it, pick "Fable 5.1" from the model picker in Claude on the web, Desktop, or Mobile. Claude Code requires version 2.1.255 or later (Fable 5 needs 2.1.170+; official support article checked September 17, 2026). Cowork needs the latest Claude Desktop. It is also available in Claude Design, Claude for Microsoft 365, and Claude Tag.
API migration — 3 breaking changes, 5 additions
If you already call claude-fable-5, check the following before swapping the model ID. The docs classify three changes as breaking:
- Forced tool use returns an error — configurations that require the model to call a specific tool no longer work on Fable 5.1. Pipelines that depend on this must be adjusted.
- Earlier models can't read its thinking blocks — passing Fable 5.1's thinking blocks in the conversation history to another model does not work.
- Editing earlier turns invalidates thinking blocks — modifying past messages in the history makes the thinking blocks after them invalid.
Five capabilities are additive, three of them in beta (a pre-release trial stage):
- Per-message effort (beta) — change the effort level partway through a conversation without invalidating the prompt cache.
- Turn-scoped system messages (beta) — system instructions that apply to a specific turn only.
- Progress updates between tool calls (beta) — readable progress via
display: "updates". - Lower cache read price — the $0.25 rate described above.
- Content provenance — provenance information attached to responses.
The docs include a dedicated migration guide covering moves from Fable 5, Opus 5, and Opus 4.8, and a page on handling classifier refusals and retrying on another Claude model. Read those on platform.claude.com before migrating production code.
Who should move to Fable 5.1
- Fable 5 users — same price, cheaper cache reads. Unless one of the three breaking changes affects you, moving is the better deal.
- Users for whom Opus 5 is enough — stay put. The official guidance keeps Opus 5 as the default and reserves Fable 5.1 for when it is needed. Fable 5.1 costs twice as much as Opus 5 on both input and output.
- Subscribers — on Max or Team premium seats, just switch in the model picker. On Pro, it runs on usage credits, so it makes sense to reserve it for work that needs it.
FAQ
Q. Can I still use Fable 5?
Yes. The docs list it as legacy but still available, with retirement not sooner than June 9, 2027. They recommend migrating to 5.1 for improved performance.
Q. Is Fable 5.1 more expensive than Fable 5?
No. Input and output stay at $10/$50, and cache reads actually dropped from $1 to $0.25.
Q. I'm on Pro — why isn't Fable 5.1 included in my limit?
On Pro and Team standard seats, both Fable 5 and 5.1 run on usage credits outside the plan limit. Inclusion applies to Max, Team premium, and Enterprise premium seats, and even there it is capped at 50% of the weekly limit.
Q. How do I get Mythos 5.1?
It is limited to Project Glasswing participants; access is requested through your Anthropic, AWS, or Google Cloud account team. The official Fable page mentions the Cyber Verification Program for cyberdefenders and says Mythos access will be included in the near future, without a specific date.
Q. Are there benchmark numbers for the improvement?
This guide covers only the specs, pricing, and availability stated in the official model and help docs. For benchmarks, refer to the system card and related materials; none are quoted here.