Fable 5 is live: Anthropic's most powerful model is now available in the playground. Chat with it free, right in your browser — no API key, no credit card. Try Fable 5 free

Introducing Fable 5 — Anthropic's new flagship

Fable 5Anthropic's most powerful AI model.

A new intelligence tier above Opus: 1M-token context, 128K output, adaptive thinking, and state-of-the-art agentic performance. Fable 5 is now live in the online playground — try it free, right in your browser.

Free credits on sign-up — no API key, no credit card.

claude-fable-5Live model

Prove that there are infinitely many primes — in two sentences.

Fable 5

Suppose only finitely many primes exist; multiply them all and add 1. That number is divisible by none of them, so it has a prime factor outside the list — contradiction. ∎

1M
token context window
128K
max output tokens
$10/$50
per 1M tokens (input / output)
Adaptive
thinking, built in

Capabilities

What makes Fable 5 different

The headline capabilities of Anthropic's new flagship model, at a glance.

Frontier intelligence

A new tier above Claude Opus 4.8 — Anthropic's most capable model to date on reasoning, coding, and knowledge work.

1M-token context window

Feed entire codebases, books, or document sets into a single conversation — one million tokens of context at standard pricing.

128K output tokens

Generate long reports, full refactors, or complete documents in one pass with up to 128K output tokens.

Adaptive thinking

The model decides when and how deeply to reason. No manual thinking budgets — depth scales automatically with task difficulty.

Long-horizon agents

State-of-the-art on autonomous multi-step work: overnight coding runs, complex refactors, and deep research that finish without correction.

High-resolution vision

Reads screenshots, charts, and documents at high resolution with pixel-accurate coordinate understanding.

API Pricing

Official Fable 5 API pricing

Official Anthropic rates compared with the rest of the Claude lineup (USD per million tokens).

ModelModel IDContextInput / MTokOutput / MTok
Fable 5NEWclaude-fable-51M$10.00$50.00
Claude Opus 4.8claude-opus-4-81M$5.00$25.00
Claude Sonnet 4.6claude-sonnet-4-61M$3.00$15.00
Claude Haiku 4.5claude-haiku-4-5200K$1.00$5.00

Official Anthropic API list prices. Our playground covers this cost through credits — no API key required.

Playground Credits

Playground credit packs

Buy credits once, spend them on any model — cheaper models cost fewer credits per message. One-time purchase, no subscription.

Starter

$9.90

A proper first test drive

  • 500 credits
  • Up to 125 playground messages
  • Credits never expire
  • Email support

Plus

$29

For regular use

  • 2,000 credits
  • Up to 500 playground messages
  • Credits never expire
  • Email support

Max

$99

For heavy users

  • 8,000 credits
  • Up to 2,000 playground messages
  • Credits never expire
  • Priority support

FAQ

Frequently asked questions

Quick answers about Anthropic's new flagship model.

Fable 5

About Fable 5: Anthropic's flagship AI model

Fable 5 (API model ID claude-fable-5, often written Fable5) is the most powerful AI model Anthropic has ever released. Launched in 2026, the Fable 5 model opens a new intelligence tier that sits above the Claude Opus family — the first of Anthropic's Mythos-class models to be generally available. Where earlier Claude models were tuned for either speed or depth, Fable 5 is built for the hardest problems: long-horizon agentic work, complex reasoning, large-scale code understanding, and enterprise-grade document analysis.

On specs, the Fable 5 model ships with a 1 million token context window — roughly 700,000 English words, or an entire codebase — and can produce up to 128K tokens of output in a single request. Thinking is always on: Fable 5 decides for itself when and how deeply to reason, with no manual thinking budgets to configure. Developers control depth through a simple effort parameter instead. The model also introduces a new tokenizer and a refreshed safety system, so prompts and token counts from older Claude models should be re-baselined.

Fable 5 pricing is straightforward: Anthropic charges $10 per million input tokens and $50 per million output tokens through the API — exactly double Claude Opus 4.8. There is no long-context premium; the full 1M window costs standard rates. If you just want to try the model without an API key, our playground sells one-time credit packs from $9.90, and new accounts get free credits on sign-up. See the full Fable 5 price breakdown on our pricing page.

How does Fable 5 compare with Claude Opus 4.8? The two models share the same API surface — adaptive thinking, no sampling parameters — but Fable 5 raises the ceiling on reasoning, coding, and agentic benchmarks, and is the better choice when output quality matters more than cost. Opus 4.8 remains the value pick for high-volume production traffic. Our detailed Fable 5 vs Opus 4.8 comparison covers pricing, context, thinking behavior, and when to choose each model.

The fastest way to understand what Fable 5 can do is to talk to it. Our online playground now runs Fable 5 right in your browser — alongside other leading models and OpenRouter Fusion — with no API key, no credit card, and free credits when you sign up. You can also put it to work immediately with our one-click AI tools: optimize a prompt, translate and polish text, or summarize a long document.

Talk to Fable 5 right now.

No API key, no waitlist. Sign in, claim your free credits, and experience Anthropic's most powerful model in your browser.