Artificial Intelligence

Claude Fable 5.2 and Mythos 5.2: What’s New

5 min read 938 words

Anthropic released two new flagship AI models today, November 9, 2026: Claude Fable 5.2 and Claude Mythos 5.2. The company calls them its most capable models yet, built for daunting technical work in coding, cybersecurity, finance, and the sciences.

Here is a simple, jargon-free breakdown of what they are, what’s new, what they cost, and what they mean for businesses.

What are Claude Fable 5.2 and Mythos 5.2?

Fable 5.2 and Mythos 5.2 are the newest models in Anthropic’s top tier. They sit above Claude Opus 5.5, which Anthropic recommends as the “daily driver” for most everyday work. Think of Opus 5.5 as the reliable all-rounder, and Fable and Mythos as the specialists you call in for the hardest problems.

The interesting part: Fable and Mythos are the same underlying model. The difference is who can access it and with which safety controls:

  • Claude Fable 5.2 — available to everyone, with strong cybersecurity and biology safeguards built in.
  • Claude Mythos 5.2 — the same model with fewer of those safeguards, available only to vetted organisations through Anthropic’s Cyber Verification Program and Life Sciences Verification Program.

What’s new in the 5.2 update?

Compared with Fable 5.1 and Mythos 5.1, the new models:

  • Generate output up to 30% faster.
  • Cost less: $8 per million input tokens and $40 per million output tokens.
  • Offer an optional fast mode that speeds up token generation up to 2.5x, at $16 per million input and $80 per million output tokens.
  • Communicate more clearly — testers reported easier-to-read, better-structured answers.
  • Show stronger results in coding, reasoning, and vision.
  • Trigger about 55% fewer unnecessary safety blocks than Fable 5.1, so legitimate work gets interrupted less often.

Benchmarks: how good are they, really?

Anthropic’s own evaluations put Fable 5.2 ahead of Claude Opus 5.5, OpenAI’s GPT-6 Astra, and Google’s Gemini 4 Carbon on most of the capabilities it tested. A few highlights:

Capability (benchmark) Fable 5.2 Opus 5.5 GPT-6 Astra Gemini 4 Carbon
Agentic coding (FrontierSWE v2) 69.9% 62.3% 65.5% 61.2%
Agentic coding (FrontierCode 2.0) 49.1% 45.5% 42.8% 41.1%
Reasoning with tools (Humanity’s Last Exam) 74.2% 67.7% 57.2% —
Computer use (OSWorld 2.1) 89.3% 87.2% 79.5% 81.5%
Robotics (RoboDojo Sim) 64.9% 39.4% 29.8% —

One nuance worth knowing: in specific AI-research areas such as chip kernel optimisation, Anthropic deliberately kept 5.2 at roughly Opus 5.5 level, to avoid the model accelerating AI development itself.

Standout real-world results

Some of the achievements Anthropic reported during pre-release testing:

  • Formalised Fermat’s Last Theorem in 30 hours. A team of about 1,000 Fable 5.2 agents translated the famous proof into machine-checkable form, with minimal human help. Anthropic’s first attempt with an earlier model took 11 days.
  • Solved two open problems in black-hole physics, both connected to uniting relativity and quantum mechanics.
  • Designed three working electronic devices — a voice-assistant hub, a USB speakerphone, and a handheld virtual pet — including circuit boards, 3D-printed enclosures, and software, each ready for manufacturing in under a day.
  • Migrated a production service from Python to Rust that now starts about 26 times faster.
  • Built a working operating system by coordinating 1,000+ subagents over eight days — complete with drivers, file storage, windows, and an installer.

Claude Fable 5.2 pricing

Mode Input (per million tokens) Output (per million tokens)
Standard $8 $40
Fast mode (up to 2.5x speed) $16 $80

For comparison, that is cheaper than the previous Fable 5.1 pricing, and Anthropic notes that at high effort, Fable 5.2 matches Opus 5.5’s best coding score at roughly half the cost per task.

Where can you use it?

  • Claude apps — Fable 5.2 is rolling out to eligible paid plans, and Anthropic is giving subscribers extra Fable usage this week as a launch promotion.
  • Claude API — available as claude-fable-5-2 on the Claude Platform.
  • Cloud platforms — Amazon Web Services, Google Cloud, and Microsoft Azure.
  • Fast mode — available in Claude Code, the Claude Platform, Google Cloud, and Microsoft Foundry.
  • Claude Mythos 5.2 — only for vetted cybersecurity and life-sciences organisations through the verification programs.

What does this mean for businesses?

Most companies do not need a frontier model for every task — and Anthropic itself suggests starting with Opus 5.5 and moving up only for hard problems. But two things in this release matter for everyday business work:

  1. Automation gets cheaper. At $8/$40 per million tokens — with 30% faster output — running AI agents for customer support, lead follow-up, research, and content operations costs noticeably less than it did last week.
  2. Agents can now run longer, harder projects. The Fermat formalisation and the Python-to-Rust migration were multi-day, multi-agent efforts. The same pattern — a coordinator model delegating to smaller ones — is becoming practical for routine business automation, not just research labs.

If you are curious where AI agents could fit in your own workflows — from chatbots to end-to-end process automation — explore our AI automation services or talk to our team.

Frequently asked questions

Are Claude Fable and Claude Mythos different models?

No. They are the same underlying model. Fable includes cybersecurity and biology safeguards and is open to everyone. Mythos removes some of those safeguards and is limited to vetted organisations in those fields.

Is Claude Fable 5.2 available on free plans?

Fable models are included with paid plans, and Anthropic is offering extra Fable usage for eligible subscribers this week. Check your plan’s usage limits in Claude settings.

How much does Claude Fable 5.2 cost via the API?

$8 per million input tokens and $40 per million output tokens. Fast mode, which generates tokens up to 2.5x faster, costs $16 per million input and $80 per million output.

Should I use Fable 5.2 or Opus 5.5?

For most coding and knowledge work, start with Opus 5.5. Move up to Fable 5.2 when a task is genuinely hard — large migrations, deep research, or multi-day agentic projects. Downgrade to Sonnet 5.5 or Haiku 5.5 for simple, high-volume tasks.

Final thoughts

Claude Fable 5.2 and Mythos 5.2 are a clear step up: faster, cheaper, and measurably better at coding, reasoning, and computer use. For developers and technical teams, they push the ceiling of what AI can do unaided. For businesses, the more practical story is that serious automation just became more affordable.

Featured image credit: Anthropic.

Share this article

Was this article helpful?

1 articles published
Free SEO Audit — No fluff, just results. Get Yours →