Claude Fable 5
FrontierLast verified: September 25, 2026· Anthropic Claude Fable 5 pricing
Previous-generation Fable, still served; superseded by Fable 5.1 at the same price. Thinking is always on; 1M-token context window.
- Input / 1M
- $10.00
- Output / 1M
- $50.00
- Context
- 1M tokens
- Provider
- Anthropic
Pricing verified September 25, 2026. Prices change frequently. Always confirm against the provider’s official pricing page before relying on these figures for budgeting. Official pricing →
When to choose Claude Fable 5
This is the top of the table and it prices like it: $10 per million in and $50 out, the only row here that costs more than $5 to read a million tokens. That premium earns its place on exactly one kind of work — the long-horizon run that either finishes or does not, where a cheaper model burns an afternoon and still fails. It carries a 1M-token window and a 128K output ceiling, so capacity is rarely the constraint. Reach for it by exception, from a route you have already tried a tier down, and keep it off anything routine.
Claude Fable 5 sits close enough to Claude Opus 4.8 on price that the two are usually decided together. That comparison is written once, on the models table, rather than argued twice from opposite sides.
What Claude Fable 5 costs in practice
The arithmetic is unforgiving at this tier. A month of 50M input and 10M output tokens costs $1,000 here, against $500 on Opus 4.8 and $180 on GPT-4.1 — the same traffic at more than five times the mid-tier price. One request that fills the 1M window costs $10 of input before Fable 5 has written a word. None of that is an argument against using it. It is an argument for knowing precisely which routes are on it, because frontier spend nobody can attribute to a route is the first thing to look for when a bill surprises you.
| Scenario | Input / mo | Output / mo | Est. cost / mo |
|---|---|---|---|
| Prototype | 2M | 0.5M | $45 |
| Growing product | 50M | 10M | $1,000 |
| At scale | 500M | 100M | $10,000 |
Plug in your own numbers with the cost calculator.
Where Claude Fable 5 fits in a routing setup
In a routed setup this is the escalation target rather than the default. The workable arrangement is Opus 4.8 carrying the frontier work and Fable 5 taking only the runs Opus demonstrably fails, which means keeping a record of those failures rather than a hunch. Because the escalation doubles the cost of whatever it touches — $1,000 against $500 on the reference month — the gate in front of it deserves as much engineering attention as the model behind it.
Cheaper alternatives to Claude Fable 5
The models immediately below Claude Fable 5 on blended cost, nearest first. A swap only pays if quality holds on the routes you move — see the LLM cost optimization playbook for how to prove that before you move them.
Frequently asked questions
How much does Claude Fable 5 cost?
Claude Fable 5 bills $10.00 per million input tokens and $50.00 per million output tokens — $20.00 per million blended at a three-to-one input:output split. A workload of 50M input and 10M output tokens per month works out at about $1,000.
What is Claude Fable 5's context window?
Claude Fable 5 has a 1M-token context window, of which up to 128K tokens can be the reply. One request that fills the whole window costs about $10.00 at the input rate, before the reply is billed.
Where does Claude Fable 5 sit on price?
At $20.00 per million blended — three input tokens for every output token — Claude Fable 5 is every one of the other 18 models tracked here is cheaper. Output bills at 5x the input rate, and the nearest cheaper model on this table is Claude Opus 4.8 at $10.00 blended.
Related concepts: context window, model routing, prompt caching, evals.