Claude Fable 5.1 cuts prices, but Pro users still pay extra: what it means for you

What happened

On September 1, 2026, Anthropic released Claude Fable 5.1 and Claude Mythos 5.1. They are the same model with different safeguards. Fable 5.1 is generally available in the Claude apps, Claude Code, the API, AWS and the other clouds. Mythos 5.1 is restricted to trusted-access programs for cybersecurity and life-sciences work, and to participants in Project Glasswing.

The launch post leads with three fixes to customer complaints. Price: Fable 5.1 costs an estimated 25 percent less than Fable 5 for typical workloads and up to about 45 percent less for long agentic jobs. Data retention: a new Enterprise Frontier Safeguards system, arriving in phases this fall, keeps customer data on infrastructure the customer controls. Safeguards: the cyber classifier now blocks 60 percent fewer benign requests, and Fable 5.1 may be used to find software vulnerabilities, though not to write exploits.

The performance gap over Fable 5 is wide. On Terminal-Bench-Science, Anthropic reports 52.6 percent against 24.7 percent. On Terminal-Bench 4.0 it reports 55.8 percent against 42.0. On Humanity's Last Exam without tools, 60.9 percent against 57.8.

Why it matters

The price cut is real but narrow. Anthropic's pricing page shows Fable 5.1 at the same base rates as Fable 5: $10 per million input tokens and $50 per million output. What changed is the price of cache reads, the tokens the model re-reads from earlier in a session. Those fell from $1 to $0.25 per million. Long coding and research sessions re-read their context constantly, which is why the saving grows with the length of the job. A five-turn chat barely notices; an hour-long Claude Code session notices a lot.

Two other things changed on the same page. Fable 5.1 has effort levels, and Anthropic says the Low and Medium settings match or beat Fable 5 at much lower cost. The default is High in Claude Code and Medium in Claude Cowork and the Claude app. And Anthropic quietly canceled a planned increase for Sonnet 5, its mid-tier model: the $2 and $10 per million introductory price, which was due to rise to $3 and $15 on September 1, is now the standard price. For anyone paying by the token, that is the bigger cut.

The context is competitive. Two days after Fable 5.1 shipped, OpenAI put GPT-6 Astra on every paid ChatGPT plan, including the $20 Plus tier. On September 19, Reuters reported that Anthropic is weighing another model release before its IPO.

What changes for a normal user

This is where the plan table matters more than the launch post. Anthropic's help center is explicit: Fable 5 and Fable 5.1 are treated identically on every plan, and a promotion that let subscribers use Fable 5 within their limits ended on July 19. Fable 5.1 was never part of it.

On the free Claude plan you do not get Fable at all. You get Sonnet 5 and Haiku 4.5, with web search, memory, file creation, voice mode and extended thinking. Nothing about that changed on September 1.

On Pro, $20 a month or $17 on an annual plan, Fable 5.1 is available but not included. It runs on usage credits, which you buy separately and spend per token at the API rates above. Unlike the July move of Fable 5, there is no one-time credit this time. Your top included model on Pro remains Opus 5, which is still very good, and Pro now bundles Claude Code, Claude Cowork, Claude Design and Claude in Chrome.

On Max, from $100 a month for 5x Pro usage or $200 for 20x, Fable 5.1 is a standard part of the plan. You can spend up to half of your weekly limit on Fable models at no extra cost. They also draw down that limit faster than Opus, so a heavy Claude Code user on Max will hit the Fable cap and then get switched to Opus or asked to buy credits.

What to do

• Free user: nothing to do. Sonnet 5 is the model you had last month. Fable-class improvements tend to reach Sonnet with the next release.
• Pro user who writes and researches: stay on Opus 5. Try Fable 5.1 with a small credit top-up only for a specific long job, and set effort to Medium.
• Pro user who codes daily: price the jump to Max 5x at $100. If you already buy credits most weeks, the included Fable allowance likely pays for itself.
• Developer: the cache-read cut and the Sonnet 5 price hold are both worth a re-quote on your monthly bill. See our model tracker for the full rate table.
• Deciding between Claude and ChatGPT at $20: read the chatbot comparison first. Claude's prose and long-document work are the reasons to pick it; the top model is not, on Pro.

Quick answers

Is Claude Fable 5.1 included in Claude Pro?

No. On the $20 Pro plan, Fable 5.1 runs only on pay-as-you-go usage credits, and there is no one-time free credit this time. Pro includes Opus 5, Sonnet 5 and Haiku 4.5. Max plans, from $100 a month, can spend up to half their weekly limit on Fable models at no extra cost.

Is Claude Fable 5.1 cheaper than Fable 5?

Per token, the base prices are unchanged at $10 per million input and $50 per million output. The cut is on cache reads, which drop from $1 to $0.25 per million tokens. Anthropic says that works out to about 25 percent less for typical workloads and up to 45 percent less for long agentic jobs.

Can I use Claude Fable 5.1 for free?

Not in the Claude app. The free plan runs Sonnet 5 and Haiku 4.5, and Fable models are not offered on it as of September 2026. Free users get Fable's improvements only indirectly, when Anthropic moves them into later Sonnet releases.

More from the deck

• 
  Tool cards

  Claude

  The writer's model, rated and priced.

• 
  Tool cards

  Cursor

  The editor where Fable-class models earn their keep.

• 
  Guides

  Free tiers and deals

  What every assistant gives away, and when to pay.