QuietlyChanged
PricingAnthropic ClaudeNotable

Fable 5.2 API pricing: $40 output, 20% cheaper than Fable 5.1 (Nov 10, 2026)

Anthropic's pricing page now lists Fable 5.2 at $40 per million output tokens, down from $50 for Fable 5.1, with prompt caching also 20% cheaper. Fable 5.1 moves to the legacy list.

The wording

Pricing · Featured API modelFirst seen Nov 10 · From our snapshot diff
Fable deleted: 5.1 inserted: 5.2 Next generation intelligence for long-running agents
Pricing · Fable output tokensFirst seen Nov 10 · From our snapshot diff
Output deleted: $50 inserted: $40 / MTok
Pricing · Fast modeFirst seen Nov 10 · From our snapshot diff
Get up to 2.5x faster speeds with fast mode for Opus 5.5 inserted: and Fable 5.2 at 2x standard pricing.
struck removed underlined added

What changed

Anthropic's pricing page changed overnight. The featured API model card that read "Fable 5.1" on Nov 9 reads "Fable 5.2" on Nov 10, with the same description: "Next generation intelligence for long-running agents."

Change 1: Lower prices on the new model

Fable 5.1's card listed prompt caching at $0.25 per million tokens (MTok) to read and $12.50 to write, and output at $50 / MTok. Fable 5.2's card lists caching at $0.20 to read and $10 to write, and output at $40 / MTok. Each of those is 20% lower. Our snapshot of the Fable 5.2 card did not capture an input-token figure, so we are not reporting one; Fable 5.1's input rate was $10 / MTok.

Change 2: Fable 5.1 moves to the older-models list

Fable 5.1 now appears further down the page, alongside Haiku 4.5, still at its old rates: $0.25 cache read, $12.50 cache write, $10 input and $50 output per MTok.

Change 3: Fast mode covers Fable 5.2

The fast mode line changed from "Get up to 2.5x faster speeds with fast mode for Opus 5.5 at 2x standard pricing" to "...for Opus 5.5 and Fable 5.2 at 2x standard pricing."

What you can do: API customers still on Fable 5.1 pay the higher rates until they switch model versions.

What stays the same: The consumer plan cards still say Fable is included on Max and Ultra "up to half your weekly limit" and is "paid with usage credits" on Pro.

Who it affects

Developers and companies paying for Anthropic's top Fable model through the Claude API, especially long-running agent workloads that produce a lot of output or reuse cached prompts

Source

Detected on the Anthropic Claude pricing page: https://www.anthropic.com/pricing. Always check the source; this summary is not legal or financial advice.

Full diff from our snapshots +23 / −10 lines
· · ·(opens in new tab)Fable 5.1Fable 5.2Next generation intelligence for long-running agents· · ·Read$0.25/ MTokWrite$12.50/ MTokInput$0.20/ MTokWrite$10/ MTokOutput$50Input/ MTokOutput$40/ MTok· · ·Learn moreGet up to 2.5x faster speeds with fast mode for Opus 5.5 at 2x standard pricing.Get up to 2.5x faster speeds with fast mode for Opus 5.5 and Fable 5.2 at 2x standard pricing.Learn more· · ·(opens in new tab)Fable 5.1Prompt cachingRead$0.25/ MTokWrite$12.50/ MTokInput$10/ MTokOutput$50/ MTokHaiku 4.5

Provenance: detected in our Nov 10 snapshot of the Anthropic Claude pricing page, compared against Nov 9.

Published by QuietlyChanged: pages tracked automatically, each change reviewed by a person before publishing. About · Contact & corrections

Keep reading

Full Anthropic Claude change history · All price changes · Report for November 10, 2026 · All reports