Anthropic
Frontier

Claude Fable 5

Frontier

Anthropic's most powerful model, a new tier above Opus. Fable 5 keeps the 1M-token context window and multimodal input (text, image, file), and targets the hardest long-horizon agentic, coding and reasoning work. Priced at $10/$50 per MTok with $1.00 cached input.

Claude Fable 5 is a frontier AI model from Anthropic. It costs $10.00 per million input tokens and $50.00 per million output tokens (blended $22.00/M), with a 1,000,000-token context window.

RatesUSD per million tokens. Blend is 70% input, 30% output, our calculation.

INPUT
$10.00/M
per million input tokens
OUTPUT
$50.00/M
per million output tokens
BLENDED 70/30
$22.00/M
unchanged since 3 May
CONTEXT
1,000,000
tokens

Hand-maintained from Anthropic's published pricing; no check date on record.

What it is good at
  • Most capable Claude to date
  • State-of-the-art long-horizon agentic execution
  • 1M-token context
  • Text, image & file input
  • Faithful to long, complex instructions
Typical use cases
  • Hardest agentic coding & codebase-wide refactors
  • Deep multi-step research
  • Overnight autonomous agent runs
  • Complex multi-tool workflows

Benchmarks

No published benchmark scores tracked for this model yet. Frontier and reasoning models from major providers have scores; smaller models and inference-host variants typically inherit the underlying open-weights score.

More from Anthropic

See all 18 →

Frequently asked questions

How much does Claude Fable 5 cost?

Claude Fable 5 costs $10.00 per million input tokens and $50.00 per million output tokens, for a blended reference rate of $22.00 per million tokens.

What is Claude Fable 5's context window?

Claude Fable 5 supports up to 1,000,000 tokens of context in a single request.

What is Claude Fable 5 best for?

Claude Fable 5 is well suited to Most capable Claude to date, State-of-the-art long-horizon agentic execution and 1M-token context.

Who makes Claude Fable 5?

Claude Fable 5 is developed and served by Anthropic.

Compared with

Terms used on this page