Claude Fable 5.2: $8 / $40, Cheaper Than Fable 5.1 (What It Costs and What Changes)
Short answer. Claude Fable 5.2, launched on November 9, 2026, costs $8 per million input tokens and $40 per million output tokens. Claude Fable 5.1 costs $10 / $50, 1.25× more on every line. A request of 10,000 input and 1,000 output tokens costs $0.12 on Fable 5.2 and $0.15 on Fable 5.1.
In plain words. Anthropic’s newest top model is cheaper than the one before it. Sending it 10,000 tokens and getting 1,000 back costs 12 cents instead of 15. A token is a short piece of text, about three-quarters of a word. The older model is still on sale.
What Anthropic announced
Anthropic’s platform release notes carry an entry dated November 9, 2026: “We’ve launched Claude Fable 5.2 ( claude-fable-5-2 ), built for demanding reasoning and long-horizon agentic work.” The same entry gives the price and the comparison in one sentence: “They cost $8 USD per MTok of input and $40 USD per MTok of output (Claude Fable 5.1 and Claude Mythos 5.1 cost $10 USD per MTok of input and $50 USD per MTok of output).”
“They” are two models. The other one is Claude Mythos 5.2, which the same entry describes as offering “the same capabilities only to organizations verified through Anthropic’s verification programs”.
Anthropic’s general pricing page, read on November 10, 2026, puts Fable 5.2 at the top of its “Model pricing” table: “Claude Fable 5.2 For demanding reasoning and long-horizon agentic work | $8 / MTok | $40 / MTok | $10 / MTok | $16 / MTok | $0.20 / MTok |”, the columns being input, output, five-minute cache write, one-hour cache write and cache read. Fable 5.1 moved to “Additional models” at its old prices. The new row is on Claude Fable 5.2’s page.
Fable 5.2 against Fable 5.1 and Opus 5.5
Every figure below is per million tokens, in US dollars, as the pricing page prints it. The last line is one request of 10,000 input tokens and 1,000 output tokens, with no cache and no Batch. The two ratio columns divide the larger price by the smaller one.
| Fable 5.2 | Fable 5.1 | Opus 5.5 | Fable 5.1 ÷ Fable 5.2 | Fable 5.2 ÷ Opus 5.5 | |
|---|---|---|---|---|---|
| Input | $8.00 | $10.00 | $4.00 | 1.25× | 2× |
| Output | $40.00 | $50.00 | $20.00 | 1.25× | 2× |
| Cache read | $0.20 | $0.25 | $0.20 | 1.25× | 1× |
| Cache write, 5 minutes | $10.00 | $12.50 | $5.00 | 1.25× | 2× |
| Cache write, 1 hour | $16.00 | $20.00 | $8.00 | 1.25× | 2× |
| Batch input / output | $4.00 / $20.00 | $5.00 / $25.00 | $2.00 / $10.00 | 1.25× | 2× |
| 10,000 in + 1,000 out | $0.12 | $0.15 | $0.06 | 1.25× | 2× |
Read the two ratio columns in their own direction. Fable 5.1 costs 1.25× what Fable 5.2 costs on every line, so the move from 5.1 to 5.2 is a price decrease. Fable 5.2 costs 2× what Claude Opus 5.5 costs on every line but one: the cache read, which is $0.20 on both.
The request line is plain arithmetic on the list prices. On Fable 5.2, 10,000 input tokens at $8 per million are $0.08 and 1,000 output tokens at $40 per million are $0.04, $0.12 in all. The same request is $0.10 plus $0.05 on Fable 5.1, and $0.04 plus $0.02 on Opus 5.5.
The cache read is 2.5% of the input price
Anthropic’s usual cache read is a tenth of the input price. The pricing page lists the exceptions on one line: “Cache read (hit) | 0.1x base input price (0.025x on Claude Fable 5.2, Claude Mythos 5.2, Claude Fable 5.1, and Claude Mythos 5.1; 0.05x on Claude Opus 5.5 and Claude Sonnet 5.5)”. It spells the figures out in the paragraph below: “On Claude Fable 5.2, Claude Mythos 5.2, Claude Fable 5.1, and Claude Mythos 5.1, a cache hit costs 2.5 percent of the standard input price ($0.20 USD per million tokens on Claude Fable 5.2 and Claude Mythos 5.2, $0.25 USD on Claude Fable 5.1 and Claude Mythos 5.1).”
So the ratio stays at 0.025× from 5.1 to 5.2, and the cache read falls with the input price, from $0.25 to $0.20. That is how Fable 5.2 ends up with the same cache read as Opus 5.5, whose input costs half as much: $0.20 is 0.025× of $8 and 0.05× of $4. Our cache price survey of September 16 covers the rule and the exceptions across providers.
Fast mode is a second price, not the list
The pricing page has a separate “Fast mode pricing” table, with the row “Claude Fable 5.2 | $16 USD / MTok | $80 USD / MTok |”. The release notes describe it in the same entry of November 9: “Fast mode delivers up to 2.5x higher output tokens per second, at $16 USD per MTok of input and $80 USD per MTok of output.” That is twice the standard price, and the fast mode page, read the same day, adds that “Fast mode is not available with the Batch API”.
AI Signal publishes the standard price, $8 / $40, as the model’s price. The $16 / $80 figure is listed on the model’s page as a second price list, with the date it was read, and is never the figure used in a table, a ranking or the calculator.
Context, output and lifecycle
Fable 5.2’s model page, read on November 10, prints “Context window 1M tokens”, “Max output 128K tokens”, “Released November 9, 2026” and “Retirement Not sooner than November 9, 2027”. Its status is “Status Active (latest)”. Anthropic’s deprecations page lists it as “claude-fable-5-2 | Active | N/A | Not sooner than November 9, 2027 |”.
Those are the same context window and maximum output as Fable 5.1, and the same as Opus 5.5, which the model page’s own comparison table prints as “Claude Opus 5.5 | 1M | 128K | $4 / $20 |”.
Claude Mythos 5.2
Claude Mythos 5.2 has its own model page, which says: “Claude Mythos 5.2 is available only to organizations verified through Anthropic’s verification programs, such as the Cyber Verification Program”, and, after a link, “It has the same specifications and pricing as Claude Fable 5.2.” Its status reads “Status Active (verification required)”, and the pricing page lists it under “Additional models” at the same five figures: “Claude Mythos 5.2 | $8 / MTok | $40 / MTok | $10 / MTok | $16 / MTok | $0.20 / MTok |”.
Because it is not open to every account, AI Signal keeps its page out of search results; its prices are still in the table and the data files.
Fable 5.1 is still on sale
Nothing on Anthropic’s pages retires Fable 5.1. Its model page, read on November 10, now says “Status Active (legacy)” and carries a link reading “Migrate to Claude Fable 5.2”. The deprecations page keeps it active, with the same retirement floor as before: “claude-fable-5-1 | Active | N/A | Not sooner than September 1, 2027 |”. Legacy is not the same as deprecated on Anthropic’s pages; only the deprecations page gives shutdown dates, and it gives none for Fable 5.1.
Its price did not change. The pricing page still lists “Claude Fable 5.1 | $10 / MTok | $50 / MTok | $12.50 / MTok | $20 / MTok | $0.25 / MTok |”. The same applies to Claude Mythos 5.1, also at $10 / $50.
What the release notes say changes when switching
The release notes of November 9 list behaviour differences alongside the price. Three of them, quoted as Anthropic writes them:
- “Claude Fable 5.2 and Claude Mythos 5.2 enforce the prefix check for every account, whatever its age.”
- “Both models require 30-day data retention, like Claude Fable 5.1 and Claude Mythos 5.1. They aren’t available under zero data retention unless expressly authorized by Anthropic.”
- “Thinking blocks that Claude Fable 5.2 and Claude Mythos 5.2 produce work only in the account that produced them, or in an account linked to it.”
Fable 5.2’s model page adds one more under “Good to know”: “Forced tool use ( tool_choice types any and tool ) returns a 400 error.”
In the Claude apps
Anthropic’s Claude pricing page, claude.com/pricing, read on November 10, 2026, names Fable without a version number. Pro’s card lists “Fable, paid with usage credits”. Max’s card lists “Fable included, up to half your weekly limit”, and the page explains below: “Both plans include Fable, our top model, for up to half your weekly limit”. The same page’s API section now prints “Fable 5.2” with “Input $8 / MTok” and “Output $40 / MTok”. The plans themselves, Pro, Max at $100 and Ultra at $200, are covered in our post on Claude Ultra and on the Claude plans page.
Every Anthropic model with its current price is on Claude API pricing, and Fable 5.2 and Mythos 5.2 are the newest entries on new models.
How we checked
- The new-models scan read Anthropic’s pricing page on the morning of November 10, 2026 and listed two names it did not have: Claude Fable 5.2 and Claude Mythos 5.2.
- The editor then read by script, the same day, the pricing page, the model pages of Fable 5.2, Mythos 5.2 and Fable 5.1, the deprecations page, the fast mode page, the release notes (entry of November 9) and claude.com/pricing. Every quote above is literal in those saved readings, which the post’s test checks.
- The rows for Fable 5.2 and Mythos 5.2 were added from those readings, with each figure’s quote and date in the row’s note. The weekly price detector, which reads the same pricing page every Monday, will check them from November 16.
- The arithmetic of the comparison table uses only the list prices above. Nothing is estimated, rounded or converted.
Frequently asked questions
How much does Claude Fable 5.2 cost?
$8 per million input tokens and $40 per million output tokens on Anthropic's pricing page, read on November 10, 2026, with cache reads at $0.20, five-minute cache writes at $10 and one-hour writes at $16. A request of 10,000 input and 1,000 output tokens costs $0.12.
Is Claude Fable 5.2 cheaper than Fable 5.1?
Yes: Fable 5.1 lists at $10 / $50, 1.25× the $8 / $40 of Fable 5.2, and the same 1.25× holds on every cache and Batch price. The release notes of November 9, 2026 give both pairs of figures.
Is Claude Fable 5.1 being retired?
Not on Anthropic's pages of November 10, 2026: its model page says 'Status Active (legacy)' and the deprecations page lists it 'Active' with retirement 'Not sooner than September 1, 2027'.
What is Claude Mythos 5.2?
A model with the same capabilities, specifications and $8 / $40 price as Fable 5.2, offered only to organizations verified through Anthropic's verification programs, such as the Cyber Verification Program.
How much is Fable 5.2 fast mode?
$16 per million input tokens and $80 output, twice the standard price, on the pricing page and the fast mode page read on November 10, 2026. It is a second price list, not the price shown for the model.
Drafted with AI assistance from the cited source and reviewed by Álvaro Lucero Rufino, who is responsible for what stays published here. Automated publishing was discontinued in September 2026; posts that no longer met that standard were removed rather than left up. This is original explanatory writing, not a reproduction — read the source above for the full story.