Anthropic: Claude Fable 5.1

Claude model for demanding reasoning and long-horizon agentic work

claude-fable-5-1

AvailableAnthropicChatCached input 98% less

Input / Output-70%: $10.00 / $50.00$3.00 / $15.00
Context: 1M
Released: Sep 1, 2026
Max output: 128K
Modalities: VisionChat
Capabilities: Tool usePrompt CacheReasoning

About Claude Fable 5.1

Claude Fable 5.1 is Anthropic's top Claude model as of September 2026, aimed at demanding reasoning and long-horizon agentic work. It is the same model as Mythos 5.1 with stricter safeguards. Against Fable 5 it scores higher on Anthropic's reported benchmarks, costs less on typical workloads, and cuts false positives from its cybersecurity and biology classifiers.

Where it works well

Anthropic reports a large jump over Fable 5 on Terminal-Bench 4.0, its agentic terminal-coding benchmark.
Anthropic's launch post reports results on CursorBench 3.2.0 and on Humanity's Last Exam without tools, ahead of Fable 5.
Safeguards flag far fewer harmless cybersecurity requests than Fable 5 did, and false positives on benign biology and medical questions drop sharply.
Adaptive thinking is always on with a high default effort, built for long agent runs that need sustained reasoning.

When to choose another model

It is the slowest tier in Anthropic's lineup, so latency-sensitive chat is better served by Sonnet 5.5 or Haiku 4.5.
Safeguards still block exploit generation and restrict some biology work, which can interrupt legitimate security research.
Anthropic itself suggests Opus 5.5 first, and Fable 5.1 when your evaluations at higher effort still fall short.

Getting started

Create API key
Create a key in Console, then use it with every model on the platform.

Send your first request

Copy the example for your language and run it against the endpoint.

Text to textMessages API

POST/v1/messages

API format:

curl https://api.tokenlab.sh/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: sk-xxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-fable-5-1",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "Hello!"}
    ]
  }'

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Default spec

per 1M tokens

Official price: Input $10.00 / Output $50.00 / Cache read $0.25 / Cache write $12.50
Verified price: Input $3.00 / Output $15.00 / Cache read $0.075 / Cache write $3.75
Discount: -70%

Prompt cache pricing

Cache read

Official price: $0.25
Verified price: $0.075
Discount: -70%

Cache write

Official price: $12.50
Verified price: $3.75
Discount: -70%

Tool fees

Charged per call, only when the model uses the tool.

Web search

Official price: $0.01/search
Verified price: $0.003/search
Discount: -70%

	Official priceper 1M tokens	Verified priceper 1M tokens	Discount
Default spec	Input $10.00 / Output $50.00 / Cache read $0.25 / Cache write $12.50	Input $3.00 / Output $15.00 / Cache read $0.075 / Cache write $3.75	-70%
Prompt cache pricing
Cache read	$0.25	$0.075	-70%
Cache write	$12.50	$3.75	-70%
Tool feesCharged per call, only when the model uses the tool.
Web search	$0.01/search	$0.003/search	-70%

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours

Requests
Success rate
P95 latency
Total tokens

Model performance

Last updated Oct 17, 2026, 12:29 PM UTC

30-day success rate: 76.2%
7-day median latency: 3.2sn=524
7-day P95 latency: 29.6sn=524
30-day requests: 2K+
Last active: 41 minutes ago

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Claude Fable 5.1 in Console with a prompt ready to edit or send.

Help me try claude-fable-5-1 with a short message at /v1/messages. Show the reply, latency, and cost.

Use cases

Best for

Reasoning
Vision

Long-horizon agents
Let an agent work for hours across a codebase or data room, planning, calling tools, and self-checking before it reports back.
Scientific analysis
Support computational biology and research work such as protein engineering analysis, GPU kernel tuning, and data exploration, within the access rules for sensitive domains.
Defensive security review
Find vulnerabilities in your own code and propose fixes, knowing that exploit generation is blocked.

Prompt examples

Review this repository for vulnerabilities, rank them by exploitability, and write a patch plan for each.

Reproduce the analysis in this notebook from the raw data, flag any step that does not hold up, and rerun it corrected.

Work through this 400-ticket backlog, group related tickets, and open a pull request for the five that share a root cause.

This model has conditional pricing. Monthly token totals alone cannot produce a reliable estimate; use the detailed pricing for the request specification, cache, and applicable time window.

FAQ

What is Claude Fable 5.1 best for?

Demanding reasoning and long-horizon agentic work: multi-hour coding runs, scientific analysis, and complex knowledge tasks. Anthropic calls it its most advanced model for coding and knowledge work, and recommends it when evaluations on Opus 5.5 at higher effort still fall short.

How does Fable 5.1 differ from Fable 5?

Anthropic reports higher scores on Terminal-Bench, Humanity's Last Exam, and other evaluations, lower cost on typical workloads, and more precise safeguards with far fewer false positives. It also has a later knowledge cutoff, so it knows more about recent software.

How is Fable 5.1 related to Mythos 5.1?

Anthropic states they are the same model with different safeguards. Mythos 5.1 has more permissive safeguards for vetted professionals in cybersecurity and life sciences, while Fable 5.1 blocks exploit generation and restricts some biology requests.

Can I turn thinking off on Fable 5.1?

No. Adaptive thinking is always on, and forced tool use returns an error. You control depth with the effort parameter instead. Anthropic's docs list these as breaking changes when migrating from earlier Claude models.

How much does Claude Fable 5.1 cost?

On TokenLab, Claude Fable 5.1 costs Input $3.00 / Output $15.00 per 1M tokens. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

What are the context window and output limit of Claude Fable 5.1?

Claude Fable 5.1 accepts up to 1,000,000 tokens of context and returns up to 128,000 tokens in one response.

Which endpoint should Claude Fable 5.1 use?

Use https://api.tokenlab.sh/v1/messages for Claude Fable 5.1. The request example below shows the matching code shape.

Which operations does Claude Fable 5.1 support?

Claude Fable 5.1 supports Text to text. Select an operation above to see its endpoint and request example.