Claude Fable 5.1 is Anthropic's most capable generally available model, built for long, demanding coding and knowledge work that runs for hours. It's on the Pro, Max, Team and Enterprise plans, not Free, and on the Claude API it costs $10 per million input tokens and $50 per million output tokens, as of October 2026.
Fable sits above Opus in Anthropic's line-up. That doesn't mean you should switch to it for everything. Anthropic's own developer docs tell most people to start with Opus 5.5 and only move to Fable when a job needs more. Here's what Fable is, how it compares and how to use it well, from Anthropic's model page, developer docs and help centre, checked as of October 2026.
Fable 5.1 next to Anthropic's other models
This comes from Anthropic's models overview. API prices are per million tokens and the figures are in US dollars.
| Fable 5.1 | Opus 5.5 | Sonnet 5.5 | |
|---|---|---|---|
| Best for (Anthropic's words) | Demanding reasoning and long-horizon agentic work | Long-running agentic coding and knowledge work | The best combination of speed and intelligence |
| Speed | Slower | Moderate | Fast |
| API input / output (as of October 2026) | $10 / $50 | $4 / $20 | $2 / $10 |
| Thinking | Always on | Always on | Adaptive |
| Default effort | High | Medium | High |
| Context window | 1M tokens | 1M tokens | 1M tokens |
| Retirement | Not sooner than 1 September 2027 | Not sooner than 22 September 2027 | Not sooner than 28 September 2027 |
Fable vs Opus: which should you use?
Anthropic's guidance is plain: start with Opus 5.5 for most workloads, and use Fable 5.1 for demanding reasoning and long-horizon agentic work, or when Opus at a higher effort setting still falls short. In other words, Fable is the model you reach for after you've tried turning Opus's effort up.
The practical difference is the length and difficulty of the job. Anthropic describes Fable as built for work that takes hours and spans many apps: working through a backlog in Cowork, picking up requests in Slack through Claude Tag (in beta), operating a browser, or running unattended as a managed agent. It plans the work, uses the tools it needs, recovers when a step fails and keeps you posted. For a quick email, a blog outline or a spreadsheet formula, it's overkill.
How to switch to Fable in Claude
- Open a new or existing chat on a Pro, Max, Team or Enterprise plan.
- Click the model name next to the send button.
- Choose Claude Fable 5.1. If you can't see it, click More models.
- Pick an effort level. Fable's default is High; Anthropic says Low and Medium stretch your usage further, while Extra high and Max are for the hardest, longest tasks.
- Send your prompt. You can change the model or effort part way through a chat, and the change applies from Claude's next reply.
On an Enterprise plan, an administrator can hide models or effort levels for your role, so a missing option isn't always a bug.
Things Anthropic says you should know first
- Thinking is always on with Fable 5.1, at every effort level, including on the API. You can't switch it off.
- Higher effort uses more of your usage limit, so a long Fable session at Max effort will run you out faster.
- Fable has safeguards for cybersecurity and biology. Many flagged queries are routed to less capable models instead, and you aren't charged Fable prices for those rerouted requests.
- Using Fable requires 30-day data retention for safety monitoring by default. Eligible Enterprise customers have other options.
- Fable understands diagrams, charts and tables inside files and PDFs, which helps with finance, legal and analytics documents.
What Fable does well, in Anthropic's words
- Ambitious coding projects: features that span a whole codebase, code review, performance work and multi-day autonomous sessions.
- Writing its own tests to check its work, and using vision to check outputs against the original design or goal.
- Multi-stage knowledge work with little oversight, from deep research and analysis to deliverables ready for your review.
- Fixing the root cause of a problem rather than the symptom, and avoiding shortcuts that only look easy.
The pattern in that list is hand-off. Anthropic pitches Fable as the model you give a large project to and then review, rather than one you supervise line by line. That's useful if you've got a messy job, like reconciling a year of reports or rebuilding a website section, that you'd normally hand to a contractor.
Fable on the Claude API
Developers call it as claude-fable-5-1 on the Claude Platform, and it's also on Amazon Web Services, Google Cloud and Microsoft Foundry. Anthropic's model page says cache reads now cost $0.25 per million tokens, and US-only inference is available at 1.1x the normal input and output price. Compare that with Opus 5.5 at $4 input and $20 output per million tokens: Fable costs two and a half times as much per token, so it pays to send it only the work that needs it.
If you'd like help choosing which models your business should use, see my AI consultancy for small businesses.
A simple way to split work between models
Here's how I'd set it up for a small business, using Anthropic's own descriptions of each model:
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
| Job | Model to start with |
|---|---|
| Quick replies, summaries, sorting messages | Haiku 5.5 |
| Everyday writing, analysis and well-scoped tasks | Sonnet 5.5 |
| Complex projects, long documents, coding | Opus 5.5 |
| Hours-long agent work or problems Opus can't crack | Fable 5.1 |
If you're weighing the middle two, my comparison of Claude Opus and Sonnet goes deeper, and this guide to every Claude plan shows which plan gets you which model.
My verdict
Fable 5.1 is impressive, but it's a specialist. Most business owners will get better value keeping Opus 5.5 or Sonnet 5.5 as their everyday model and saving Fable for the big, gnarly job that would otherwise eat a whole afternoon. If you're on Pro, try it on one hard task this week and see if the result is worth the extra usage.
Not every job needs the biggest model, and Claude Sonnet vs Haiku shows where the smaller ones are enough.
Frequently asked questions
What is Claude Fable?
Claude Fable 5.1 is Anthropic's most capable generally available model, built for demanding reasoning and long-running agentic work.
Is Claude Fable free?
No. Anthropic says Fable 5.1 is available to Pro, Max, Team and Enterprise users, not the Free plan.
Is Fable better than Opus?
Fable is more capable, but Anthropic recommends starting with Opus 5.5 for most work and using Fable when Opus still falls short.
How much does Claude Fable cost on the API?
As of October 2026, $10 per million input tokens and $50 per million output tokens, with cache reads at $0.25 per million.
Can I turn off thinking on Claude Fable?
No. Anthropic says thinking is always on with Fable 5.1 at every effort level.
