Claude Fable 5.2 in Claude Code: Alias, Effort, Fast Mode
Claude Code 2.1.321 added Claude Fable 5.2 (claude-fable-5-2) and made it the model the fable alias resolves to, except in Claude apps gateway sessions, where fable still means Fable 5. The changelog lists it at $8 per million input tokens and $40 per million output, with cache reads at $0.20 per million and a 1M-token context window. It does not become anyone's default: the docs say no Fable model is the account-type default on any plan or provider, so you choose it yourself.
This page covers what is different from Fable 5.1 in the places that affect cost and setup. Verified November 10, 2026 against the 2.1.321 and 2.1.322 changelog entries and the Claude Code model configuration and fast mode pages. We did not run a Fable 5.2 session (this was written without API access), so everything below is documented behavior, not measured behavior.
How to select it
claude --model fable
Or run /model fable in a session. On any provider other than a Claude apps gateway, fable now resolves to Fable 5.2. Fable 5.1 and Fable 5 are still available by model ID (/model claude-fable-5-1, /model claude-fable-5) on the Anthropic API, and on other providers by your provider's model ID or by pinning with ANTHROPIC_DEFAULT_FABLE_MODEL.
Fable 5.2 needs Claude Code 2.1.321 or later. If you are on an older version and the request fails, the errors page points to "Claude Code does not support this model", and claude update is the fix. As of our check, npm latest is 2.1.322.
What happens to a saved Fable 5.1 pick
On Pro, Max, Team, and Enterprise plans, a claude-fable-5-1 model saved in your user settings moves to the fable alias, so you land on Fable 5.2.
- Pro or Max: the move happens the first time you run 2.1.321 or later. The startup model line can show
(auto-updated)once. - Team or Enterprise: the move waits for your organization's settings. It can come at a later startup or not at all, applies from the next session, and shows no
(auto-updated)line. Run/modelto see where you are.
Console accounts, third-party providers, and project, local, or managed settings keep claude-fable-5-1. If you want a team to stay on 5.1, put the model ID in project settings, which the docs say is not rewritten.
Effort starts at medium
Per the model configuration page, Fable 5.2 defaults to medium effort, the same as Opus 5.5, Sonnet 5.5, and Haiku 5.5. After the alias move, sessions start at 5.2's default unless you have set a level for it. The supported levels for Fable models are low, medium, high, xhigh, and max, and thinking cannot be turned off on any Fable model. Our effort levels page has the full precedence order. If you relied on Fable 5.1's default, set the level explicitly:
claude --model fable --effort high
We have not measured how output length differs between 5.1 and 5.2 at the same level, so do not read the default change as a cost change in either direction until you have looked at your own /usage.
Fast mode on Fable 5.2
The fast mode page lists Fable 5.2 as supported (up to 2.5x faster at a higher per-token price) and gives these rates:
| Model | Fast mode input (MTok) | Fast mode output (MTok) |
|---|---|---|
| Fable 5.2 | $16 | $80 |
| Opus 5.5 | $8 | $40 |
| Opus 5 | $10 | $50 |
Against the changelog's standard Fable 5.2 rate of $8/$40, the fast rate is twice the standard rate on both input and output (arithmetic on the two published figures). Three details are specific to Fable 5.2:
"fastMode": truealone does nothing on Fable 5.2. You have to run/fastwhile using Fable 5.2.- Switching to Fable 5.2 turns fast mode off unless you turned it on while using Fable 5.2. It persists across sessions only if you enabled it there, unless you set per-session opt-in.
- Subscription plans pay for fast mode from usage credits, not included limits, per the fast mode page.
Fast and standard speed keep separate caches, and the first time you turn fast mode on mid-conversation you pay the full fast-mode uncached input price on the whole context. Turn it on at the start of the session if you will use it. Fable 5.2 fast mode also depends on your organization's usage tier. 2.1.322 added fast mode on Vertex in the global, us and eu regions, falling back to standard speed if Vertex refuses it. The related 2.1.321 fix says that when fast serving is rate-limited, overloaded or failing, Claude Code now tries standard speed on the same model first instead of losing replies or retrying for minutes.
Usage credits can bill Fable
Depending on plan and seat tier, Fable usage can bill to usage credits instead of your plan's included limits. The /model picker then shows "Requires usage credits" on the Fable row, and interactive sessions can show a consent prompt. In claude -p and in Agent SDK applications that do not show the prompt, Claude Code never asks: a Fable request that would bill to usage credits is billed without asking. For scheduled jobs, that is the line to read before you set --model fable in a script.
Gateways and allowlists
- Claude apps gateway:
fablestays on Fable 5. A gateway not configured to serveclaude-fable-5-2rejects it, so on one that does serve it, select it with/model claude-fable-5-2. availableModels: a version prefix matches later IDs.claude-fable-5permits Fable 5, 5.1, and 5.2, whileclaude-fable-5-1permits 5.1 only. An allowlist written for 5.1 will not let 5.2 through.- Third-party providers: pin
ANTHROPIC_DEFAULT_FABLE_MODELto your provider's Fable model ID. The docs say this also makes Claude Code treat it as a Fable model for automatic fallback.
Safety fallback still applies
Requests that Fable's safety classifiers flag, most often cybersecurity and biology, trigger automatic model fallback: biology-flagged requests re-run on Opus 5 and cybersecurity-flagged requests on Opus 4.8. The docs note that penetration testing, CTF work, and biology-adjacent codebases trigger fallback often, sometimes on the first request. The 2.1.321 changelog also fixed the fallback model sometimes declining a request or mentioning a stopped response when the first model was blocked before replying.
FAQ
Is Fable 5.2 the default model in Claude Code?
No. The model configuration page says no Fable model is the account-type default on any plan or provider. Choose it with /model fable or --model fable.
Why did my Fable 5.1 session change?
On Pro, Max, Team, and Enterprise, a claude-fable-5-1 value saved in user settings moves to the fable alias from 2.1.321. Project, local, and managed settings are left alone.
Does Fable 5.2 cost more than Opus 5.5?
Yes, at standard speed. Fable 5.2 is $8/$40 per million tokens (2.1.321 changelog). Our Opus 5.5 page recorded Opus 5.5 at $4/$20 when it shipped, so Fable 5.2 is twice that. Recheck the pricing reference before budgeting, since rates can change.
How do I turn fast mode on for Fable 5.2?
Switch to Fable 5.2, then run /fast. Editing fastMode in settings does not enable it for Fable 5.2 by itself.