Claude Fable 5.2 Cuts the Price to $8 per Million Tokens, but Pro Users Get One Week of It
Anthropic cut Fable 5.2 to $8 per million input tokens and $40 per million output tokens, but Pro and Team standard seats only get Fable included for one week. Max and Ultra users get one to three resets, Mythos stays vetted, and the Fermat proof claim has no independent review.
Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on 9 November 2026, priced at $8 per million input tokens and $40 per million output tokens. That is 20% below Fable 5.1, which the Claude pricing docs list at $10 and $50 per million tokens.
Anthropic says output is up to 30% faster than Fable 5.1 and Mythos 5.1, with a 1 million token context window and 128,000 token maximum output. An optional fast mode doubles the price to $16 and $80 per million tokens and speeds generation by up to 2.5 times.
The catch is not the model. It is the access. Anthropic is including Fable models for Pro plans and standard Team seats for one week only, capped at 50% of weekly usage limits. After 16 November 2026, Fable usage on those plans goes back to usage credits.
Quick answer: Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens, 20% less than Fable 5.1. It is available now on the Claude Platform, AWS, Google Cloud and Microsoft Azure. Pro and Team standard seats get Fable included until 16 November 2026, up to 50% of weekly limits, while Max, Ultra and premium Team seats get one to three Fable resets. Mythos 5.2 stays limited to vetted cybersecurity and life sciences organizations.
What Anthropic changed, and what it did not
Fable and Mythos are the same underlying model. The difference is packaging. Fable ships with cybersecurity and biology safeguards and is available to all users. Mythos is reserved for organizations accepted into Anthropic’s Cyber Verification Program or Life Sciences Verification Program.
The capability gains are concentrated in agentic coding, business workflows and long running science tasks. Anthropic’s own benchmark table puts Fable 5.2 ahead of Fable 5.1 and Opus 5.5 on every row below, though one rival still beats it on AutomationBench.
| Benchmark | Fable 5.2 | Fable 5.1 | Opus 5.5 | GPT-6 Astra | Gemini 4 Carbon |
|---|---|---|---|---|---|
| FrontierCode 2.0 | 49.1% | 32.4% | 45.5% | 42.8% | 41.1% |
| FrontierSWE v2 | 69.9% | 56.3% | 62.3% | 65.5% | 61.2% |
| AutomationBench | 51.4% | 31.4% | 42.5% | 41.4% | 53.9% |
| Price per million tokens | $8 input, $40 output | $10 input, $50 output | Not compared here | Not compared here | Not compared here |
Read the table with care. These are Anthropic’s numbers, on benchmarks Anthropic chose. Gemini 4 Carbon still beats Fable 5.2 on AutomationBench, 53.9% to 51.4%, so the new model is not a clean sweep. Anthropic says Fable 5.2 at High effort scored about level with Opus 5.5 at Max effort, for roughly half the cost per task.
That cost per task claim is worth testing. A cheaper model that thinks longer can still cost more.
The price cut is real, and smaller than the launch framing suggests
A 20% cut is meaningful at scale. It is not a change in category. Fable 5.2 is still a premium model, and fast mode puts it back above Fable 5.1’s old list price. Meta’s Muse Wonder is cheaper still, at $4 input and $20 output per million tokens.
| Configuration | Input per million tokens | Output per million tokens | Approximate INR at 94.05 |
|---|---|---|---|
| Fable 5.2 standard | $8 | $40 | Rs 752 / Rs 3,762 |
| Fable 5.2 fast mode | $16 | $80 | Rs 1,505 / Rs 7,524 |
| Fable 5.1 standard | $10 | $50 | Rs 941 / Rs 4,703 |
Anthropic recommends that most coding and knowledge work start on Claude Opus 5.5, then move up to Fable or Mythos only when the task is hard enough to justify the price. That guidance matters more than the benchmark table. Fable 5.2 is designed for ambitious, open ended technical work, not for every chat request.
Anthropic also says it deliberately capped the model’s ability to accelerate AI research itself, including accelerator kernel optimization and model architecture work, where Fable 5.2 performs at about the level of Opus 5.5. It says agentic coding and hardware development were not affected.
The Fermat claim is the real headline, and it is unaudited
The most striking claim in the launch post is not a benchmark score. Anthropic says a team of 1,000 Fable 5.2 agents fully formalized the proof of Fermat’s Last Theorem in 30 hours, starting from minimal prompts, with minimal human intervention and no mathematical references.
Anthropic says the run used one third of the action tokens of its first formalization and produced a proof seven times shorter. The comparison is stark: the first formalization, by an unreleased research version of Claude, took 11 days with a purpose built harness and human guidance. With the same 1,000 agent setup, Opus 5.5 completed only 26% of the formalization in 18 hours and barely progressed after 30.
There are two caveats. Anthropic’s own chart says the comparison is a single run per model. And the company has not published an independent mathematician’s review of the result. Formal verification in Lean is stronger evidence than a model’s prose, but it still depends on the formal statement matching the theorem people think was proved.
Anthropic also says Fable 5.2 solved two open problems in the physics of black holes, both tied to holographic duality. Those claims have not been peer reviewed as of publication. Treat them as company claims, not settled science.
What Pro and Max users actually get
Anthropic’s launch post says Fable is included for up to 50% of weekly usage limits for Pro plans and standard Team seats. The support article is more precise, and more limited.
- Pro plans and standard Team seats: all Fable models included from 9 November 2026 through 16 November 2026 at 11:59 pm PT, up to 50% of weekly usage limits. After that, Fable usage is billed to usage credits if they are enabled.
- Max plans, Ultra plans, premium Team seats and premium seats on legacy seat based Enterprise plans: one to three Fable limit resets, depending on how long the account has been a paid subscriber.
- Excluded: Free, standard seats on legacy seat based Enterprise plans, self serve Enterprise plans and usage based Enterprise plans.
The billing cutoff matters. Resets go only to accounts with paid billing in place before 9 November 2026 at 10:00 am PT. A subscription that started before that date but is still in a free trial gets resets only after it converts to paid.
So the answer to the question many users are asking, whether Fable 5 is included in Max plans, is yes for this launch offer, through resets. It is not a promise that Fable stays inside the standard Max limit after the promotion ends.
Confirmed versus not confirmed
| Claim | Status |
|---|---|
| Fable 5.2 and Mythos 5.2 launched on 9 November 2026 | Confirmed by Anthropic |
| Standard price is $8 input and $40 output per million tokens | Confirmed by Anthropic and the Claude pricing docs |
| Fable 5.1 was priced at $10 input and $50 output | Confirmed by the Claude pricing docs |
| Context window is 1M tokens and max output is 128K tokens | Confirmed by the model overview page |
| Output is up to 30% faster than Fable 5.1 | Company claim, methodology not detailed in the post |
| Pro and Team standard seats get Fable for one week only | Confirmed by the Claude Help Center |
| 1,000 agents formalized Fermat’s Last Theorem in 30 hours | Company claim, single run, no independent review published with the launch |
| Fable 5.2 solved two open black hole problems | Unconfirmed, not peer reviewed |
| Fable 5.2 beats every competitor on every benchmark | Not confirmed, Gemini 4 Carbon leads AutomationBench in Anthropic’s own table |
| Fable stays included in Max after the launch promotion | Not confirmed, Anthropic has not said so |
What it means for Indian developers
At 94.05 rupees to the dollar, Fable 5.2 costs about Rs 752 per million input tokens and Rs 3,762 per million output tokens. Fast mode doubles both numbers. That is still expensive for bulk work, which is why the plan details matter more than the benchmark gains.
A developer on Claude Pro gets one week of Fable access at no extra cost, capped at half the weekly limit, then pays usage credits. A team on standard seats gets the same window. A Max or Ultra subscriber gets one to three resets instead, which rewards long tenure rather than high usage.
The practical move is to treat Fable 5.2 as an escalation model. Use Opus 5.5 or Sonnet 5.5 for routine work, and reserve Fable for migrations, hard debugging, formal methods and research tasks where a single better answer saves hours.
Data handling is the other India relevant detail. Anthropic says Fable 5.2 and Mythos 5.2 rely on safety monitoring that can review a window of usage data, with 30 day retention for ordinary customers. Zero data retention is available to eligible enterprise customers, and Enterprise Frontier Safeguards runs monitoring inside a customer’s infrastructure. Teams handling customer data should confirm retention terms before routing production traffic through the model.
Anthropic lists availability on AWS, Google Cloud and Microsoft Azure, but makes no India specific hosting or data residency claim in the launch post.
FAQ
How much does Claude Fable 5.2 cost?
Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens. Fast mode costs $16 per million input and $80 per million output. Fable 5.1 cost $10 and $50, so the standard price is 20% lower.
Is Fable 5.2 included in Claude Pro and Max?
On Pro and standard Team seats, all Fable models are included from 9 November to 16 November 2026, up to 50% of weekly usage limits. After that, Fable usage goes to usage credits. Max, Ultra and premium Team seats get one to three Fable resets instead. The offer is a launch promotion, not a permanent plan change.
How is Fable 5.2 different from Mythos 5.2?
They are the same underlying model with different safeguards and access. Fable has cybersecurity and biology classifiers and is available to all users. Mythos is available only to organizations accepted into Anthropic’s Cyber Verification Program or Life Sciences Verification Program.
Did Claude really prove Fermat’s Last Theorem?
Anthropic says a team of 1,000 Fable 5.2 agents formalized the proof in Lean in 30 hours. That is a company claim from a single run per model, and no independent mathematician’s review was published with the launch. Formalization is strong evidence, but the formal statement still needs to match the theorem.
What is the context window of Fable 5.2?
Fable 5.2 has a 1 million token context window and a maximum output of 128,000 tokens. Its reliable knowledge cutoff is June 2026, and Anthropic says it will not retire the model sooner than 9 November 2027.
Sources
- Anthropic, “Introducing Claude Fable 5.2 and Claude Mythos 5.2,” 9 November 2026. https://www.anthropic.com/claude-fable-and-mythos-5-2
- Claude Platform Docs, “Pricing,” accessed 9 November 2026. https://platform.claude.com/docs/en/about-claude/pricing
- Claude Platform Docs, “Claude Fable 5.2,” accessed 9 November 2026. https://platform.claude.com/docs/en/models/fable-5-2/overview
- Claude Help Center, “Claude Fable 5.2 launch offers,” accessed 9 November 2026. https://support.claude.com/en/articles/17506494
- Anthropic, “Claude Fable 5.2 and Mythos 5.2 system card,” 9 November 2026. https://www.anthropic.com/document/claude-fable-5-2-mythos-5-2-system-card
- Anthropic, “Formalizing Fermat’s Last Theorem,” accessed 9 November 2026. https://www.anthropic.com/research/formalizing-fermats-last-theorem