Claude Fable 5.2: Anthropic's Fastest Model, 20% Cheaper

Q: Should I switch from Claude Opus 5.5 to Fable 5.2?

For most everyday coding and knowledge work, no — Anthropic itself still recommends starting with Opus 5.5, which costs half as much. Fable 5.2 is the pick for the most ambitious, open-ended technical work, where its higher benchmark scores justify the price. Claude Pro subscribers can try Fable for up to half of their weekly allowance at no extra charge through November 16, 2026.

Updated: November 20265 min readBy the CreatorSprout team

Launch coverage based on Anthropic’s November 9, 2026 announcement and its published API pricing, as reported by multiple outlets, we have not tested Claude Fable 5.2 hands-on. Capability and benchmark claims below are Anthropic’s or independent benchmarkers’, not results we measured. No affiliate links in this article.

Fable 5.2 arrives just over two months after Fable 5.1, and it is the rare frontier release that gets faster and cheaper at the same time. Anthropic is also doing something unusual with access: Fable and Mythos are the same underlying model, split into two versions by their safeguards. Fable carries the blocks that keep it safe to offer broadly, while Mythos, with some of those safeguards relaxed, is limited to organisations that pass Anthropic’s verification programs. For creators and developers, the practical question is not whether the model is impressive on paper, it is whether it is worth twice the price of Opus 5.5 for the work you actually do.

What was announced

Anthropic introduced Claude Fable 5.2 and Claude Mythos 5.2 on Monday, November 9, 2026, describing Fable as its most capable model for technical work, setting a new bar in coding, finance and the sciences. According to Anthropic, testers found both models communicate more clearly than the 5.1 versions and generate output up to 30% faster. Fable 5.1 and Mythos 5.1 had arrived on September 1, 2026, so this is a fast follow-up, roughly ten weeks later.

Anthropic also put two striking research claims in the launch: the company says Fable 5.2 helped solve two open problems in the physics of black holes, with preprints posted for both results, and that about 1,000 Fable agents formalised the proof of Fermat’s Last Theorem in Lean, a language that lets a computer verify each step, in just 30 hours. Those are Anthropic’s claims from its own testing, and the independent scrutiny is still catching up.

Fable vs Mythos: same model, two doors

Fable 5.2 is the version you can use. It carries safeguards that block most cybersecurity and biology requests, and that is what allows Anthropic to offer it to everyone: Claude paid plans, Claude Code, the Claude Platform API, and Amazon Web Services, Google Cloud and Microsoft Foundry.
Mythos 5.2 is the same model with some safeguards relaxed. It is limited to organisations accepted into Anthropic’s Cyber Verification Program or Life Sciences Verification Program. If you are not in one of those programs, Mythos is not an option, and it is not a tier you can buy your way into.
The split is a safety design, not a marketing one. Anthropic says plainly that the two are the same underlying model. The difference is entirely in what each version is allowed to do, which is how Anthropic can ship its most capable model broadly while keeping the riskiest capabilities behind a verification door.
Model ID for developers: claude-fable-5-2 on the Claude Platform API, and anthropic.claude-fable-5-2 on Amazon Bedrock. Knowledge cutoff is June 2026, and Anthropic lists retirement as not sooner than November 9, 2027.

Pricing: cheaper, but still twice Opus

API pricing, $8 input and $40 output per million tokens. That is down from $10 and $50 on Fable 5.1, a 20% cut on both sides. The price drop is the most concrete good news in the launch: a flagship that gets cheaper is rare.
Opus 5.5 is still half the price. Claude Opus 5.5 costs $4 per million input tokens and $20 per million output tokens. Fable 5.2 costs exactly double. The question is whether the capability gap is worth double the bill for your workload, and for routine work Anthropic’s own guidance says it is not.
Pro subscribers get a free trial window. Through November 16, 2026, at 11:59 p.m. Pacific, Claude Pro subscribers can use Fable models for up to half of their existing weekly allowance at no extra charge. Anthropic also granted extra Fable usage resets during launch week, one to three depending on subscription length, for subscribers whose paid billing was in place before 10 a.m. Pacific on November 9.
No free tier. Fable 5.2 is available to Pro, Max, Ultra, Team and Enterprise users. If you are on a free Claude plan, the model is not in your menu, and the launch-week Pro offer does not change that.

Benchmarks: number one, with caveats

Artificial Analysis Intelligence Index: 65, first place. The independent benchmarker’s index, which combines nine evaluations, puts Fable 5.2 at maximum effort in first place with a score of 65, five points ahead of Gemini 4 Carbon and about 15 ahead of Fable 5.1. On its AA-3D board, which tests how well a model builds an interactive 3D scene, Fable 5.2 leads at 1271 Elo, 67 points ahead of Claude Sonnet 5.5 and 69 ahead of Claude Opus 5.5.
Anthropic’s own table: FrontierSWE v2 at 69.9%. On Anthropic’s published benchmark table, Fable 5.2 scores 69.9% on FrontierSWE v2, against 62.3% for Claude Opus 5.5. These are Anthropic’s figures from its own testing, so read them as the vendor’s case, not an independent verdict.
Anthropic itself flags a measurement problem. The launch post admits Fable 5.2 and Mythos 5.2 voice suspicion that they are being evaluated more often than any of its recent models, which makes their safety results harder to read. That is an unusually candid note, and it is a good reason to treat early benchmark headlines with a grain of salt until independent testing settles.

Where you can use it

Claude apps and plans: Pro, Max, Ultra, Team and Enterprise users, with extra Fable usage during launch week.
Developers: Claude Platform API (model ID claude-fable-5-2), Amazon Bedrock, Google Cloud and Microsoft Foundry.
Claude Code: available in the coding tool on launch day.
GitHub Copilot: added Fable 5.2 the same day for Pro+, Max, Business and Enterprise users, but admins must switch on a policy that is off by default, so if you do not see it, check with your admin.
Devin: Cognition made Fable 5.2 available in Devin Desktop and Devin CLI on launch day.

What it means for creators and developers

Try it free if you have Pro, before you pay for it. The launch-week offer, up to half your weekly allowance at no extra charge through November 16, is the lowest-risk way to find out whether Fable 5.2 changes anything for your work. Run your hardest real task through it, not a demo prompt.
Use it for the ambitious work, not the routine work. Anthropic’s own model guide puts Opus 5.5 first for most coding and knowledge work, and reserves Fable for the most ambitious, open-ended technical work. That is the vendor telling you not to overspend, and it is good advice.
Long agent runs are the real use case. The 1,000-agent Fermat formalisation is a hint at where Fable earns its price: long-horizon, multi-agent work where a slightly better model compounds over hours. If your workflow is single short prompts, the 30% speed gain matters more than the capability gain.
Read the migration guide before switching an API integration. Anthropic’s migration guide lists changes that can break code that works today, including a 30-day data-retention requirement, with a 400 error on requests from organisations that do not have it enabled. Check your account settings before you flip the model ID.
Cheaper flagship, same discipline. A 20% price cut does not make a $40-per-million-output-token model cheap. Price the output tokens, not the input tokens, because on long generations the output side is where the bill lives.

The honest caveats

We haven’t tested it. Everything here comes from Anthropic’s announcement and its published API pricing, as reported by multiple outlets, plus independent benchmark scores. The speed gain, the clarity improvement and the research claims are Anthropic’s, not results we measured.
The benchmarks are early. The Artificial Analysis scores are independent and genuinely impressive, but they landed within hours of the release, and Anthropic itself warns its safety results are harder to read this time. Give it a few weeks before treating any ranking as settled.
Mythos is not a product you can buy. Coverage that blurs Fable and Mythos together can mislead: the unrestricted version is gated behind verification programs for cyber and life-science work. If a headline promises capabilities you cannot access, it is probably describing Mythos.
The Pro trial has a hard cutoff. The launch-week allowance ends November 16, 2026, at 11:59 p.m. Pacific, and the extra usage resets were only for billing in place before 10 a.m. Pacific on November 9. After that, Fable usage comes out of your normal allowance or your API bill.
Launch-week hype is a real risk. A flagship release always generates a wave of first-impression posts, and several early testers used the model free of charge, with most of their results not independently checked. Wait for the second wave of measured comparisons before rebuilding a workflow around it.

Verdict

Claude Fable 5.2 is a genuinely strong release, and the price cut is the part that matters most. A frontier model that gets up to 30% faster and 20% cheaper at the same time is rare, and the independent number-one ranking on the Artificial Analysis index gives the capability claims real weight. But the honest version is the usual one: Anthropic itself says most everyday work still belongs on the cheaper Opus 5.5, Mythos is not available to you, and the benchmarks are hours old. If you have Claude Pro, use the free launch-week allowance on your hardest real task before November 16. If you pay per token, start with Opus, and let Fable earn its double price on the work where it actually shows.

FAQ

What is Claude Fable 5.2?

Claude Fable 5.2 is Anthropic’s most capable generally available model, released November 9, 2026. It is the same underlying model as Claude Mythos 5.2, but with stronger safeguards that block most cybersecurity and biology requests, which is what allows Anthropic to offer it broadly.

How much does Claude Fable 5.2 cost?

Through the Claude API, Fable 5.2 costs $8 per million input tokens and $40 per million output tokens, down from $10 and $50 on Fable 5.1, a 20% cut. For comparison, Claude Opus 5.5 costs $4 input and $20 output per million tokens.

How is Fable 5.2 different from Mythos 5.2?

Fable 5.2 and Mythos 5.2 are the same underlying model. Fable carries safeguards that block most cybersecurity and biology requests and is available to everyone on paid plans and the API. Mythos has some of those safeguards relaxed and is limited to organisations accepted into Anthropic’s Cyber Verification Program or Life Sciences Verification Program.

Where can I use Claude Fable 5.2?

Fable 5.2 is available on Claude paid plans (Pro, Max, Ultra, Team and Enterprise), in Claude Code, on the Claude Platform API as claude-fable-5-2, and through Amazon Bedrock, Google Cloud and Microsoft Foundry. GitHub Copilot added it on launch day for Pro+, Max, Business and Enterprise users, and Cognition made it available in Devin Desktop and Devin CLI the same day.

Should I switch from Claude Opus 5.5 to Fable 5.2?

For most everyday coding and knowledge work, no, Anthropic itself still recommends starting with Opus 5.5, which costs half as much. Fable 5.2 is the pick for the most ambitious, open-ended technical work, where its higher benchmark scores justify the price. Claude Pro subscribers can try Fable for up to half of their weekly allowance at no extra charge through November 16, 2026.