Released June 9, 2026
What is Fable 5? The Model That Compresses Months of Work into Days
A Mythos-class model, made safe for general use.
Fable 5 is the Mythos-class model Anthropic made safe for general use in June 2026. Its successor, Claude Fable 5.1, arrived on September 1, 2026. This page covers what Fable 5 is, what it does well, what it costs, and how to start working with it.
Your first Fable 5 project. Minutes, not hours.
Fable 5 is an Anthropic model released June 9, 2026, which Anthropic describes as a Mythos-class model made safe for general use. It is the same underlying model as Claude Mythos 5, which has some safeguards lifted and is limited to verified organizations. At launch Anthropic called it state-of-the-art on nearly all tested benchmarks, and its API price is $10 per million input tokens and $50 per million output tokens. Claude Fable 5.1 succeeded it on September 1, 2026, and Fable 5 is now a legacy model that is still available.
For every Claude release with its date, see the Claude Model Timeline.
Published · Last reviewed
Free lesson · no account neededMeet ClaudeWhy Claude?6 minStart the free lesson| Model | Released | Status | Context window | Max output | Reliable knowledge cutoff |
|---|---|---|---|---|---|
| Claude Fable 5 | June 9, 2026 | Legacy, still available | 1M tokens | 128K tokens | January 2026 |
| Claude Fable 5.1 | September 1, 2026 | Current | 1M tokens | 128K tokens | June 2026 |
| Claude Opus 5.5 | September 22, 2026 | Current | 1M tokens | 128K tokens | June 2026 |
Capabilities
What Fable 5 does well
The defining trait of Fable 5 is that performance scales with task length and complexity. The longer and harder the work, the more the model separates itself.
Software engineering
State-of-the-art on software engineering tasks. Stripe used Fable 5 to compress months of engineering work into days.
Knowledge work
State-of-the-art performance on knowledge work, with performance that scales as tasks get longer and more complex.
Vision tasks
Extracts data from scientific figures and rebuilds web apps from screenshots. Vision is a first-class capability, not an add-on.
Scientific research
State-of-the-art on scientific research tasks, from reading figures to long-context reasoning across large bodies of material.
Long-duration autonomous work
Performance scales with task length and complexity. Built for long-running autonomous work, not just quick exchanges.
Safety routing built in
Sensitive queries auto-route: cybersecurity questions go to Claude Opus 4.8, with bio/chem and distillation protections in place. Triggers in under 5% of sessions.
Long-Horizon Work
Built for tasks that take hours, not seconds
Most model comparisons focus on short benchmarks. Fable 5 is designed for the opposite end of the spectrum: long-duration autonomous work and long-context reasoning across millions of tokens. That means whole codebases, full research corpora, and multi-step projects that run without constant supervision.
The practical effect shows up in how teams use it. Stripe compressed months of engineering work into days. On the vision side, Fable 5 can extract data from scientific figures and rebuild web apps from screenshots, which turns visual artifacts into working starting points instead of dead ends.
This is why learning to work with Fable 5 matters: the model rewards people who can scope, delegate, and verify long-running work. Those are skills, and they can be learned.
Pricing
What Fable 5 costs
API Pricing
Subscriber Access
Launch window ended June 22, 2026
From launch through June 22, 2026, Fable 5 was included at no extra cost on Pro, Max, Team, and seat-based Enterprise plans. Anthropic said that from June 23, using it on those plans would require usage credits.
Safety
How Fable 5 handles sensitive queries
Before Fable 5, the only Mythos-class model was Claude Mythos Preview, offered only to Project Glasswing partners for defensive cybersecurity work. To make Fable 5 generally available, Anthropic launched it with safeguards. Claude Mythos 5 is the same underlying model with some of those safeguards lifted, and it is available only to verified organizations.
In practice, certain sensitive queries are automatically routed to Claude Opus 4.8: some cybersecurity questions, most biology and chemistry requests, and requests flagged as model distillation attempts. Anthropic says these safeguards trigger, on average, in less than 5% of sessions, so for the vast majority of work you will never notice them.
Learn how to actually work with Fable 5
Reading about a model and being good with it are different things. The free Fable 5 course at Nightschool AI walks you through scoping long-running tasks, working with vision inputs, and getting reliable results from a model that rewards clear delegation. Hands-on from the first lesson.
Frequently Asked Questions
What is Fable 5?
Fable 5 is an Anthropic model released on June 9, 2026. Anthropic describes it as a Mythos-class model made safe for general use. It is the same underlying model as Claude Mythos 5, which has some safeguards lifted and is available only to organizations verified through Anthropic programs. Its successor, Claude Fable 5.1, was released on September 1, 2026, and Fable 5 is now a legacy model that is still available.
What is the Mythos class?
Anthropic uses Mythos-class for models that sit above its Opus class in capability. The first, Claude Mythos Preview, was released in April 2026 only to Project Glasswing partners for defensive cybersecurity work, and Anthropic does not plan to make it generally available. Fable 5 was the first Mythos-class model Anthropic made safe for general use.
How much does Fable 5 cost?
Fable 5 API pricing is $10 per million input tokens and $50 per million output tokens. From launch through June 22, 2026, it was included on Pro, Max, Team, and seat-based Enterprise plans at no extra cost.
Is Fable 5 free?
Not anymore. From launch through June 22, 2026, Fable 5 was included on Pro, Max, Team, and seat-based Enterprise plans at no extra cost. Anthropic said that from June 23, using it on those plans would require usage credits. API pricing is $10 per million input tokens and $50 per million output tokens.
How is Fable 5 different from Claude Opus?
Fable 5 is a Mythos-class model, a class Anthropic places above Opus in capability, and its performance scales with task length and complexity, making it suited to long-duration autonomous work and long-context reasoning across millions of tokens. The two also work together: when Fable 5 receives certain sensitive cybersecurity queries, it automatically routes them to Claude Opus 4.8. Anthropic says these safeguards trigger, on average, in less than 5% of sessions.
Is Fable 5 still Anthropic's newest model?
No. Claude Fable 5.1, released September 1, 2026, is the successor to Fable 5. Anthropic's current lineup is Claude Fable 5.1, Claude Opus 5.5, Claude Sonnet 5.5, and Claude Haiku 5.5, plus Claude Mythos 5.1 for organizations verified through Anthropic's verification programs. Fable 5 is a legacy model that remains available. The Claude Model Timeline on Nightschool AI lists every release with its date.
What is Fable 5 best at?
At launch, Anthropic described Fable 5 as state-of-the-art on nearly all tested benchmarks, including software engineering, knowledge work, vision tasks, and scientific research. Its vision capabilities include extracting data from scientific figures and rebuilding web apps from screenshots. Stripe reported compressing months of engineering work into days with it.
Get hands-on with Fable 5 this week
The free course is open, and the skills carry over to Fable 5.1.