claude-fable-5.1 API — per-token pricing and code examples

Checking statusPrice version 2 · in effect since Nov 19, 2026

Price per 1M tokens

InputOfficial price $10.00$3.10
Cache readOfficial price $0.25$0.0775
Cache writeOfficial price $12.50$3.875
Cache write 1 hOfficial price $20.00$6.20
OutputOfficial price $50.00$15.50

Minimum per request$0.00002

Cache rates apply when the provider reports cache tokens.

69% off

About this model

claude-fable-5.1 is served through OpenAI Chat Completions, OpenAI Responses, Anthropic Messages. You pay $3.10 per 1M input tokens and $15.50 per 1M output tokens. That is 69% below the official price of $10.00 and $50.00. Cache reads cost $0.0775 per 1M tokens when the provider reports cached tokens. It is included in budget packs from $99.00 up.

Where Claude Fable 5.1 fits

Claude Fable 5.1 is Anthropic's latest Fable model, released on September 1, 2026, with retirement listed as not sooner than September 1, 2027 and knowledge through June 2026. Anthropic positions it "for demanding reasoning and long-horizon agentic work" and lists its latency as "Slower". Input can be text or images, output is text, the context window is 1M tokens, and a synchronous reply can run to 128K tokens.

Anthropic is also clear about when not to start with it. For most workloads it recommends beginning with Claude Opus 5.5 and moving to Fable 5.1 only when your evals on Opus 5.5 at higher effort still do not meet your requirements. In other words, Fable 5.1 is the step you take after measuring, for the hardest multi-step problems where extra quality is worth the wait.

Thinking that stays with Fable 5.1

Adaptive thinking is the only thinking mode on Fable 5.1, and it is always on. Requests that send thinking: {type: "disabled"}, between_tools or a budget_tokens value are rejected with 400. Effort defaults to high, all five levels are accepted, and Anthropic suggests stepping up to xhigh or max for the most capability-sensitive agentic and coding work. Unlike Fable 5, this model supports per-message effort as a beta.

Two rules about thinking blocks matter if you build agents or switch models mid-conversation. Anthropic states that only Claude Fable 5.1 and Claude Mythos 5.1 can read Fable 5.1's thinking blocks, so a conversation started on Fable 5.1 cannot simply be handed to another Claude model with its thinking intact. And editing earlier turns invalidates the thinking blocks, so rewrite history carefully.

Forced tool use is the main breaking change from Fable 5: it now returns an error. Replace a named tool_choice with {"type": "auto"} plus strict tool use, or ask for a JSON result through structured outputs. Anthropic also notes that some advisor tool pairings are not supported.

What it costs on Cent AI

Fable 5.1 is billed per token from your prepaid USD balance at the rates in the price box, and because thinking cannot be disabled, every call includes reasoning tokens charged at the output rate. Long agent runs at xhigh or max produce the most of them. Anthropic's own advice to test Opus 5.5 first is a sensible cost check before you move traffic here.

Fable 5.1 can cache prompts from 512 tokens. The default lifetime is five minutes and the optional lifetime is one hour; cache writes for both are billed at the input rate. Cache pricing only applies to the cached tokens Fable 5.1 reports in its usage. Since editing earlier turns invalidates thinking blocks, an append-only conversation history is also the one that keeps cache reads working.

Using it from your tools

Fable 5.1 is reachable through /v1/messages, /v1/chat/completions and /v1/responses, and Cent AI does not change your parameters silently. Anthropic documents that non-default temperature, top_p or top_k values return 400 on Fable 5.1, and that API version 2026-10-21 refuses those fields at any value. A prefilled last assistant turn is refused too.

A /v1/messages request at xhigh effort for a long planning task:

{
  "model": "claude-fable-5.1",
  "max_tokens": 16000,
  "output_config": { "effort": "xhigh" },
  "messages": [{ "role": "user", "content": "Plan the database split step by step" }]
}

The Cent AI model name keeps a dot, claude-fable-5.1, while Anthropic's API ID is written claude-fable-5-1.

Specification

Context window
—
Input and output
Text
Endpoints
/v1/chat/completions, /v1/responses, /v1/messages
Smallest pack
$99.00

Latency and error rate per model are on the status page. status.cent-ai.dev

What a request costs

Open the full calculator

Cost per request

90% is an assumption you can change, not a promise. Cache is billed only from cache tokens the model provider reports.

See how many such requests it covers.

Estimate

claude-fable-5.1
Fresh input99,000 tok
Cache read891,000 tok
Output10,000 tok

Per request$0.5309
At official pricesAt official prices $1.7127

Total$0.5309

An estimate at price version 2. Your bill comes from the usage the provider reports for each request.

Save $1.1817

Code

curl https://api.cent-ai.dev/v1/chat/completions \
  -H "Authorization: Bearer $CENT_AI_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-fable-5.1", "messages": [{"role": "user", "content": "Hello"}]}'

Set CENT_AI_KEY to a key from the dashboard. OpenAI SDKs use https://api.cent-ai.dev/v1; Anthropic SDKs use https://api.cent-ai.dev.

Price history

Version 2, in effect since Nov 19, 2026.

  1. Effective

    per 1M tokens

    Inputold $1.30 new $3.10
    Cache readold $0.13 new $0.0775
    Cache writeold $1.30 new $3.875
    Cache write 1hold $1.30 new $6.20
    Outputold $6.50 new $15.50

Compare with

All models against official prices

Questions

How much does claude-fable-5.1 cost on Cent AI?

$3.10 per 1M input tokens and $15.50 per 1M output tokens. Cache reads cost $0.0775 and cache writes $3.875 per 1M tokens. Each request is rounded up once.

Is there a minimum charge?

A successful request costs at least $0.00002. A request that ends early, because the client left or the output was cut off, pays only for its tokens.

How do I pay?

Top up a prepaid USD balance with QRIS or USDT/USDC, or buy a budget pack once. Nothing renews automatically.

What if the provider does not report usage?

The request is billed from an estimate and marked estimated on the Requests page. An estimate never contains cache tokens.

Which endpoint do I call?

/v1/chat/completions, /v1/responses, /v1/messages. The OpenAI base URL is https://api.cent-ai.dev/v1.

Try claude-fable-5.1 with one key

One prepaid balance for every model. Pay with QRIS or USDT/USDC.

Start free