ESENRUUK

Fable 5.2 special: what changes with Claude's new model and how to put it to work

November 11, 2026 · 5 min read · By the Master en Claude team · News MasterEnClaude

The short version. Anthropic launched Claude Fable 5.2 on November 9, 2026. It replaces Fable 5.1 as Anthropic's most capable model open to all customers and is built for demanding reasoning and long-horizon agent work: sessions that run for hours, multistep research and analysis carried through to a finished document, spreadsheet or deck (official model page). In the apps it is available on the Pro, Max, Ultra, Team and Enterprise plans, according to the Fable page.

What changes compared with Fable 5.1

• It costs less on the API. $8 per million input tokens and $40 per million output tokens, down from $10 and $50 for Fable 5.1. Cache reads cost $0.20 per million (official “What's new in Fable 5.2”).
• It writes faster. Anthropic reports outputs up to 30% faster than Fable 5.1, plus an optional fast mode that generates up to 2.5 times faster at a higher price (November 9 announcement).
• Better at coding, reasoning and vision. According to the same announcement, it beats Opus 5.5 on almost every capability Anthropic evaluated.
• Fewer blocks. Queries its safeguards link to cybersecurity or biology are still routed to other models (Opus 4.8 and Opus 5), but Anthropic reports about 55% fewer blocks than with Fable 5.1.
• A separate cap on Pro. On Pro and on standard Team seats, Fable can use up to half of the weekly usage limit (announcement).

The Claude lineup as of November 11, 2026

This is how the official models table looks now:

• Fable 5.2 (November 9): demanding reasoning and long-horizon agent work.
• Opus 5.5 (September 22): long-running coding and knowledge work. Still the model Anthropic recommends starting with in most cases.
• Sonnet 5.5 (September 28): the best balance of speed and intelligence.
• Haiku 5.5 (October 7): the fastest and cheapest, for high-volume tasks.

Fable 5.1 moves to the list of earlier models but is still available. Mythos 5.2 has the same capabilities as Fable 5.2 and is reserved for organizations verified through Anthropic's programs.

Three things you will notice

Anthropic lists them in its guide to prompting Fable 5.2:

• It almost always thinks before answering. Asking it to “answer directly” no longer holds it back the way it did with Fable 5.1. If a simple question takes too long, lower the effort level instead of rewording the line.
• Sometimes it stops before the end. On long jobs it may finish with a list of what is left or an offer to continue. If you want the whole job, say so in the brief: “finish everything; if something blocks you, explain what and carry on with the rest”.
• It follows style rules to the letter. A rigid sentence-length limit can make the prose choppy. Review instructions you wrote for earlier models.

If you use it through the API

• The model ID is claude-fable-5-2.
• Thinking is always on: the old budget_tokens mode returns an error, and depth is set with the effort parameter, which defaults to high.
• It does not accept temperature, top_p or top_k, or forcing a tool with tool_choice.
• Computer use goes through the new computer_toolset_20260801 tool.
• If your code edits earlier messages and sends the model's thinking back, the request fails: the history can only be appended to.

What does not change

Anthropic's advice is the same as with 5.1: start with Opus 5.5 and move to Fable when the work really calls for it. What makes a good brief does not change either: who the result is for, what material there is, what shape it should take and what to do with doubts. People who learn to explain the work do not start from zero with every new model.

Facts checked on November 11, 2026 on the Anthropic pages linked in the text. Models and their terms change: when in doubt, the official page wins.

← All news