Claude Fable 5 & Mythos 5 — Anthropic's Most Powerful Models Yet, Explained
Anthropic just changed the game
On June 9, 2026, Anthropic launched two new models: Claude Fable 5 and Claude Mythos 5. These are not incremental updates. They represent a new tier of AI capability — what Anthropic calls "Mythos-class" — and they are the most powerful models the company has ever released.
Here is everything you need to know.
What is Claude Fable 5?
Fable 5 is the general-availability model. It is available to everyone — developers, businesses, and regular users — and it is state-of-the-art on nearly every benchmark Anthropic has tested.
The name comes from the Latin fabula, meaning "that which is told" — related to the Greek mythos. The distinction between Fable and Mythos comes down to safeguards, which we will explain below.
What makes Fable 5 special?
Software engineering. Stripe tested Fable 5 on a 50-million-line Ruby codebase and reported that the model completed a full codebase-wide migration in a single day — a task that would have taken a whole engineering team over two months by hand.
Knowledge work. On Hebbia's Finance Benchmark for senior-level reasoning, Fable 5 scored highest among all models tested, with strong gains in document reasoning, chart interpretation, and analytical problem solving.
Vision. Fable 5 is the new state-of-the-art model for vision tasks. It can extract data from complex scientific figures, rebuild a web app from screenshots alone, and even beat the Pokémon FireRed video game using only raw screen pixels — no maps, no helper tools.
Long-context and memory. The model stays focused across millions of tokens and can use its own notes to improve outputs over long tasks — making it ideal for agentic, multi-step workflows.
What is Claude Mythos 5?
Mythos 5 is the same underlying model as Fable 5, but with certain safety safeguards lifted — specifically in the area of cybersecurity.
It is not publicly available. Mythos 5 is currently restricted to a small group of trusted partners through Project Glasswing, a collaboration with the US government focused on cyber defense. Anthropic plans to expand access through a broader trusted access program over time.
Fable 5 vs Mythos 5 — what is the difference?
| Feature | Claude Fable 5 | Claude Mythos 5 |
|---|---|---|
| General availability | Yes | No (restricted) |
| Cybersecurity safeguards | Active (falls back to Opus 4.8) | Lifted for trusted partners |
| Biology safeguards | Active (falls back to Opus 4.8) | Lifted for select researchers |
| Underlying model | Same | Same |
| Price | $10 / $50 per million tokens | $10 / $50 per million tokens |
The safeguards explained
Fable 5 comes with a new class of safety classifiers — separate AI systems that monitor requests in real time. When a request touches sensitive areas, it automatically falls back to Claude Opus 4.8 instead.
The three areas covered by classifiers:
- Cybersecurity — prevents the model from assisting with offensive hacking, exploit development, or cyberattacks.
- Biology and chemistry — blocks requests related to dangerous biological or chemical research.
- Distillation — detects attempts to extract Claude's capabilities to train competing models.
Anthropic says more than 95% of Fable 5 sessions involve no fallback at all. When a fallback does happen, users are notified, and they still get a high-quality response from Opus 4.8.
What this means for our vibe coding workflow
We have been using Claude Sonnet and Opus extensively in our organization for vibe coding. Fable 5 is a significant upgrade for the tasks we care about most:
- Large codebase migrations — Fable 5's ability to work autonomously across millions of lines of code is a game changer
- Complex architecture work — the model's reasoning depth on long, multi-step tasks is noticeably stronger
- Vision-based tasks — building UI from screenshots or reading complex diagrams is now far more reliable
- Long agentic sessions — Fable 5 maintains focus and uses notes to stay on track across extended Claude Code sessions
We are already integrating claude-fable-5 via the API for our most demanding projects.
Pricing and availability
Fable 5 is available now via the Claude API using the model string claude-fable-5.
Pricing: $10 per million input tokens and $50 per million output tokens — less than half the price of the previous Mythos Preview model.
Subscription plans: Fable 5 is included on Pro, Max, Team, and Enterprise plans at no extra cost through June 22, 2026. After that, usage credits will be required until capacity allows it to return as a standard plan feature.
Our take
Fable 5 is the most capable model Anthropic has made publicly available, and based on early reports from partners like Stripe and IMC, it delivers real, measurable gains on complex tasks — not just benchmark improvements.
For teams already using vibe coding, this is a meaningful upgrade. The longer and more complex the task, the more Fable 5 pulls ahead.
We will be publishing a hands-on review of Fable 5 in our vibe coding workflow soon. Stay tuned.
This post is part of our series on AI-assisted development. Read previous posts: How We Use Vibe Coding | Claude Models We Use and Why
Enjoyed this article? Let's work together.
Contact Ficiali