Models

Claude Fable 5.1: pricing, specs and how to use it in Claude Code

Published November 2, 2026 · Last checked November 2, 2026

Claude Fable 5.1 is Anthropic's most capable generally available model, released on September 1, 2026 for demanding reasoning and long-running agentic work. This post collects its official specs and prices, how it differs from Fable 5 and Opus 5.5, and how to use it in Claude Code. Specs and prices come from Anthropic's Fable 5.1 model page and pricing page, checked on November 2, 2026.

Official specs

1M tokens is roughly 2.5M Unicode characters on the current tokenizer. Claude Mythos 5.1 shares Fable 5.1's specs and prices but is available only to organizations verified through Anthropic's verification programs.

Pricing

Fable 5.1 has the lowest cache-read multiplier of any model: most models use 0.1, and Opus 5.5 and Sonnet 5.5 use 0.05. In a long Claude Code session most tokens are context read back from the cache, so the real cost gap between Fable 5.1 and Opus 5.5 is usually smaller than the 2.5 times in their list prices. Output still costs $50 per million, so output-heavy work stays expensive. Every model's prices are in Claude API pricing.

Compared with Fable 5 and Opus 5.5

Anthropic says Fable 5.1 extends Fable 5 with stronger long-running agentic coding, multistep research, and document, spreadsheet and slide work. Code that calls the API directly and moves from Fable 5 has four breaking changes (What's new in Claude Fable 5.1):

• Forced tool use returns an error.
• Only Fable 5.1 and Mythos 5.1 can read Fable 5.1's thinking blocks, so another model cannot continue the conversation with them.
• Editing earlier turns invalidates the thinking blocks after them.
• Some advisor tool pairings are no longer supported.

The additions are mid-conversation effort changes, turn-scoped system messages and readable progress updates between tool calls (all three in beta), plus the lower cache-read price and content provenance. In Claude Code, Claude Code handles these itself.

Which tasks are worth Fable 5.1

Anthropic's models overview says: when unsure, start with Opus 5.5; use Fable 5.1 for demanding reasoning and long-horizon agentic work, or when your evals on Opus 5.5 at higher effort still fall short. Claude Code's model guide calls Fable suited to tasks larger than a single sitting: it sustains long autonomous sessions, investigates before acting, and verifies its work more often.

Claude Code's documentation suggests:

• Describe the outcome, not the steps. Hand it the result you want and let it plan the path.
• Hand it ambiguous problems. Root-cause investigations, outage debugging and architecture decisions are where the extra investigation pays off.
• Skip the verification reminders. It checks its own work with less prompting.
• Size up larger tasks. Give it work you would normally split into pieces.

For small edits, single functions and questions, Sonnet 5.5 or Opus 5.5 is faster and cheaper. How to split work between the models is in Choosing a Claude model.

Using it in Claude Code

Fable is not the default model on any account type, so select it yourself. Fable 5.1 needs Claude Code v2.1.257 or later; requests from older versions fail, so run claude update first.

Anthropic's documentation says that through a gateway or relay, you select Fable 5.1 by its full model ID. In Claude Code, run:

/model claude-fable-5-1

For one session only, launch with:

claude --model claude-fable-5-1

To start every session on it, set it in ~/.claude/settings.json:

{
  "env": {
    "ANTHROPIC_BASE_URL": "https://corecode.live",
    "ANTHROPIC_AUTH_TOKEN": "sk-your-key"
  },
  "model": "claude-fable-5-1"
}

Claude Code runs Fable 5.1 at high effort by default; change it with /effort. Switching to Fable mid-session writes the whole context to the cache again at the write price on the next turn, so pick the model when the task starts.

Safety classifiers and automatic fallback

Fable models run with safety classifiers that most often flag cybersecurity and biology content. When one flags a request, Claude Code re-runs it on another model: biology on Opus 5, cybersecurity on Opus 4.8, and the session stays on the fallback model until you switch back with /model. CoreCode also serves Opus 5 and Opus 4.8, so the fallback runs as it does on the Anthropic API, and the re-run request is billed at the fallback model's price.

Using it through CoreCode

Both of CoreCode's Claude groups, Claude 满血（仅限CC）(Claude Code only) and Claude 满血（可外接）(any client), serve claude-fable-5-1, answered by official Anthropic Claude accounts. The groups are compared in Groups.

Fable costs more to serve, so its discount on CoreCode is smaller than on the other Claude models. Current prices are in the price table, synced from the console every 10 minutes. To get started, see Getting started.

FAQ

Sources

• Anthropic: Claude Fable 5.1 (specs, model IDs, release and retirement dates; checked November 2, 2026)
• Anthropic: Pricing (prices and cache multipliers)
• Anthropic: Models overview (model choice, comparative latency)
• Claude Code: Model configuration (selecting Fable, version requirement, default effort, automatic fallback)