Claude Fable vs Opus: Which Model Should You Use?

Claude Fable vs Opus compared: Fable 5.2 and Opus 5.5 prices, default effort, plan access, and when the extra cost of Fable is worth it for your work.

Josh Pocock8 min read

Fable vs Opus cover

AI Coding

Fable vs Opus

Stride AI Academy

In Claude Fable vs Opus, the short answer is: start with Opus 5.5 and move to Fable 5.2 only when Opus falls short. Fable 5.2 is Anthropic's most capable tier at $8 per million input tokens and $40 per million output tokens, exactly twice Opus 5.5's $4 and $20. Anthropic's own docs say to start with Opus 5.5 for most work and use Fable for demanding reasoning, long agent runs, or when Opus at higher effort still isn't good enough.

This guide compares the two on price, specs, effort, plan access and Claude Code, then gives a simple way to decide. Everything here was checked against Anthropic's models overview, API pricing docs, Claude's pricing page, the Help Center article on Fable models on your plan and Claude Code's model configuration docs on November 22, 2026.

Fable 5.2 vs Opus 5.5 at a glance

From the models overview and pricing docs, November 22, 2026:

Claude Fable 5.2 Claude Opus 5.5
Anthropic's description Demanding reasoning and long-horizon agentic work Long-running agentic coding and knowledge work
API model ID claude-fable-5-2 claude-opus-5-5
Input price (per MTok) $8 $4
Output price (per MTok) $40 $20
Comparative latency Moderate Moderate
Thinking Adaptive, always on Adaptive, always on
Default effort (API) high medium
Default effort (Claude Code) medium medium
Context window 1M tokens 1M tokens
Max output 128K tokens 128K tokens
Knowledge cutoff June 2026 June 2026
Default model in Claude Code Never Yes, on every paid plan

MTok means a million tokens. Both take text and image input and support vision and tool use. On paper the specs match. The differences are price, the API's default effort and where each one is meant to be used.

What Anthropic says about choosing

The models overview puts it plainly: if you're unsure, start with Opus 5.5 for most workloads. It suggests Fable 5.2 for demanding reasoning and long-horizon agentic work, or "when your evals on Claude Opus 5.5 at higher effort still fall short."

That last part matters. The docs don't tell you to jump to Fable the moment a task feels hard. They tell you to raise Opus's effort first, test it, and only switch if the results still aren't there.

Claude Code's docs describe Fable as its most capable models, "suited to tasks larger than a single sitting". They say Fable sustains long autonomous sessions, investigates before acting and checks its work more often than smaller models.

Fable vs Opus pricing

Fable costs twice as much on every line of the bill. From the pricing docs:

Cost per MTok Fable 5.2 Opus 5.5
Input $8 $4
Output $40 $20
5-minute cache write (1.25x) $10 $5
1-hour cache write (2x) $16 $8
Cache read $0.20 $0.20
Batch input / output (50% off) $4 / $20 $2 / $10
Fast mode input / output $16 / $80 $8 / $40

The cache read line is the surprise. Fable 5.2's cache reads cost 0.025x its input price and Opus 5.5's cost 0.05x, so both land at $0.20. If your work rereads a large cached prompt many times, such as a long agent session over the same codebase, the gap between the two narrows on that part of the bill.

A quick worked example

Say one task uses 2 million fresh input tokens and 300,000 output tokens:

  • Opus 5.5: 2 x $4 + 0.3 x $20 = $8 + $6 = $14
  • Fable 5.2: 2 x $8 + 0.3 x $40 = $16 + $12 = $28

Fable is only cheaper overall if it finishes the job in less than half the tokens, or saves you a rerun. For every Opus model's costs, see Claude Opus pricing, and for the full API price list, Claude API pricing.

Effort: try this before switching models

On the API, Fable 5.2 defaults to high effort and Opus 5.5 to medium. So a raw API call to Fable thinks harder by default as well as costing more per token. In Claude Code, both default to medium.

Both models accept low, medium, high, xhigh and max in Claude Code. Before you pay for Fable, try Opus at a higher effort:

Terminal
/effort high

If Opus at high or xhigh solves the task, you've saved half the per-token price. If it still misses, that's the case where Anthropic's docs point you to Fable.

Fable and Opus on Claude plans

Opus 5.5 is included in every paid plan's usage. Fable works differently, per the Fable Help Center article and the pricing page:

Plan Opus 5.5 Fable 5.2
Free ($0) No No
Pro ($20/month, or $17/month billed yearly) Included in plan limits Usage credits from the first request
Max ($100/month) Included Up to 50% of your weekly limit, then usage credits
Ultra ($200/month) Included Up to 50% of your weekly limit, then usage credits
Team standard seat Included Usage credits
Team premium seat Included Up to 50% of your weekly limit

Two things to know about the 50%:

  1. It isn't extra. Fable draws from your normal weekly pool, and it uses that pool faster than other models.
  2. At the cap you choose. Keep using Fable on usage credits, or switch to another model and stay inside your plan.

On Pro, every Fable request bills to usage credits at API rates, so Fable on Pro is pay-as-you-go. If you want Fable as part of your plan, that's Max or Ultra, and our Claude Max plan guide covers what else Max adds. For how weekly limits work in general, see Claude usage limits, and for usage credits, Claude credits.

On Team and Enterprise, an owner or admin controls which models are enabled. On Enterprise, the article says Fable 5.2 stays off until an admin turns it on, and is added seven days after launch to roles that already have a Fable model, unless an admin opts out.

Using Fable and Opus in Claude Code

Claude Code runs Opus 5.5 by default on Pro, Max, Ultra, Team and Enterprise plans and with an Anthropic API key. The model configuration docs say no Fable model is the default on any plan or provider, so you pick it yourself:

Terminal
/model fable

Or at startup:

Terminal
claude --model fable

Fable 5.2 needs Claude Code v2.1.321 or later. When your plan bills Fable to usage credits, the /model picker shows "Requires usage credits" on the Fable row. To go back to Opus:

Terminal
/model opus

Our guide to changing the model in Claude Code covers aliases, the settings key and the order Claude Code applies them in.

A practical pattern: keep Opus as your default for everyday building, and switch to Fable for the long, hard sessions, such as a large refactor or a multi-hour agent run you'll leave alone. Then switch back. Put your own hours into the Claude Code cost calculator to see what that mix costs.

When to pick each one

Pick Opus 5.5 when:

  • You're not sure. It's Anthropic's recommended starting point.
  • You're building day to day: features, fixes, reviews, client tools.
  • You're on Pro and don't want every request to cost extra.
  • Cost per task matters and Opus at higher effort gets the job done.

Pick Fable 5.2 when:

  • The task is long and you'll leave the agent running for hours.
  • It needs deep reasoning across a large codebase or many documents.
  • You've tested Opus 5.5 at higher effort and it still falls short.
  • You're on Max or Ultra and have Fable room left in your week.

If you're also weighing the cheaper models, Claude Sonnet vs Opus covers the next step down, and every current Claude model sits side by side in our model guide. For a closer look at Opus 5.5 itself, see our free Opus 5.5 guide.

Learn to build with Claude Code

The Stride AI Coding course teaches Claude Code and Codex for building client tools, written for non-programmers, with weekly live calls to review your builds. Many of the free guides cover Claude Code setups, and the Claude Code community is where members compare theirs.

FAQ

Is Claude Fable better than Opus?

Fable 5.2 is Anthropic's most capable tier, built for demanding reasoning and long agent runs. But Anthropic recommends Opus 5.5 for most work and suggests Fable only when Opus at higher effort still falls short.

How much more does Fable cost than Opus?

Twice as much per token. Fable 5.2 is $8 input and $40 output per million tokens, and Opus 5.5 is $4 and $20. Cache reads are the exception: both cost $0.20 per million tokens.

Can I use Fable on Claude Pro?

Yes, but not from your plan limits. On Pro, Fable bills to usage credits at API rates from the first request. Max and Ultra can spend up to half of their weekly limit on Fable.

Does Claude Code use Fable or Opus by default?

Opus 5.5, on every paid plan and with an API key. No plan defaults to Fable. Select it with /model fable or claude --model fable, which needs Claude Code v2.1.321 or later.

What is the difference between Fable 5.2 and Opus 5.5 specs?

Both have a 1M-token context window, 128K max output, adaptive thinking and a June 2026 knowledge cutoff. The differences are price and the API's default effort: high for Fable and medium for Opus.

Should I raise effort or switch to Fable?

Raise effort first. Anthropic's docs suggest moving to Fable when Opus 5.5 at higher effort still falls short in your tests. Run /effort high in Claude Code and see if that's enough.

Build it with us in Stride AI Academy

Four courses, weekly live calls and a community of people building and selling AI. Or start with 320+ free guides.

Keep reading

Free AI Coding guides