What is Claude Fable 5?
A plain-language guide to Claude Fable 5 — Anthropic’s most capable publicly available model, launched June 9, 2026 — what it can do, the safeguards built into it, and what it changes for businesses.
A new capability tier, made public
Claude Fable 5 is the publicly available version of Anthropic’s “Mythos”-class models — a tier the company positions above its Opus class in raw capability. It launched on June 9, 2026, and Anthropic describes it as exceeding every model it had previously made generally available, scoring state-of-the-art on nearly all of the benchmarks it tests.
Alongside it, Anthropic announced Claude Mythos 5: the same underlying model with some safeguards lifted, restricted to a small set of vetted partners. Fable 5 is the version the public can use today; Mythos 5 is the gated one. If you are reading about “Mythos” and wondering what you can actually access, the answer is Fable 5.
Coding, vision, and long-horizon work
The headline strengths are in long, autonomous tasks. On Cognition’s FrontierCode evaluation, Anthropic reports Fable 5 scores highest among frontier models even at medium effort, and Stripe described it compressing months of engineering into days. It is built to hold complex, long-horizon coding work together rather than losing the thread halfway through.
- Software engineering — reliable on long, multi-step coding tasks with a high degree of autonomy.
- Vision — Anthropic calls it the new state of the art for vision, citing the ability to read precise numbers off scientific figures and rebuild a web app’s source code from screenshots alone.
- Knowledge work and research — strong results across analysis, scientific reasoning, and structured knowledge tasks.
- Long context — stays focused across millions of tokens; with persistent file-based memory, Anthropic reports it improved on the Slay the Spire benchmark three times more than Opus 4.8 did.
Why some answers come from Opus 4.8
Fable 5 ships with classifiers that automatically route a narrow set of concerning requests — chiefly around offensive cybersecurity, and biology and chemistry — to Claude Opus 4.8 instead of answering directly. Anthropic says more than 95% of sessions involve no such fallback at all, and that external red-teaming found no universal jailbreaks across more than 1,000 hours of testing.
Availability and access
Fable 5 is available through Anthropic’s API from launch day, and is generally available in GitHub Copilot for Pro+, Max, Business, and Enterprise plans across editor modes including agent mode. Anthropic is offering it at well under half the cost of its earlier Mythos preview, which materially lowers the barrier to running frontier-grade work in production.
One detail matters for sensitive environments: in GitHub Copilot, Fable 5 requires data retention — prompts and outputs are kept for up to 30 days for safety purposes, and administrators must enable the model deliberately because it is off by default. Whether and how a model retains your data is exactly the kind of question that should be answered before, not after, you build on it.
Capability is only half the job
A more capable model raises the ceiling on what an autonomous system can do; it does not, by itself, run your business. The capability has to be wrapped in a harness that gives it the right tools, the right data, bounded authority, monitoring, and a place to run that respects where your data is allowed to live. That wrapper is where most of the real engineering — and most of the risk — actually sits.
This is the work AIMOCS does: we run models like Fable 5 as managed operators that take on a specific workflow end to end — contained, audited, and in-region where that is required — so the business gets the capability without having to manage tokens, infrastructure, or the safety plumbing around it.
What is Claude Fable 5?
Claude Fable 5 is Anthropic’s most capable publicly available model, launched June 9, 2026. It is the public version of the company’s “Mythos” class, a capability tier it positions above the Opus class, with state-of-the-art results across software engineering, vision, knowledge work, and research.
How is Claude Fable 5 different from Claude Mythos 5?
They are the same underlying model. Fable 5 is the public version with full safeguards; Mythos 5 has some safeguards lifted and is restricted to a small set of vetted partners such as cyber-defense and biology researchers. Fable 5 is what the public can use today.
Why does Claude Fable 5 sometimes answer from Opus 4.8?
Fable 5 uses classifiers that route a narrow set of sensitive requests — mainly offensive cybersecurity and biology or chemistry — to Claude Opus 4.8 instead. Anthropic says this fallback happens in fewer than 5% of sessions.
Where can I use Claude Fable 5?
It is available through Anthropic’s API from launch and is generally available in GitHub Copilot for Pro+, Max, Business, and Enterprise plans across editor modes including agent mode. Note that in Copilot it requires data retention and is off by default until an administrator enables it.
Is Claude Fable 5 safe to use with business data?
It can be, but the safe path depends on the deployment — how data is retained, where it is processed, and what the model is allowed to do. Those choices, plus bounded agent authority and monitoring, matter more than the raw benchmark score for most business use.
We don't advise on AI. We run it for you.
Proven on your data before you commit.