Anthropic launches Claude Fable 5.2 and Mythos 5.2: what's new

On November 9, 2026, Anthropic announced two new models, Claude Fable 5.2 and Claude Mythos 5.2. According to Anthropic, they are the same underlying model with different safeguards and different access rules. Fable 5.2 is widely available, while Mythos 5.2 is limited to vetted organizations.

Nearly every performance figure below comes from Anthropic's own announcement. We say so each time, because company-reported results are not independent verification.

What was announced

• Claude Fable 5.2 is available on all platforms, Anthropic says, including Amazon Web Services, Google Cloud and Microsoft Azure. Developers use the API model string claude-fable-5-2.
• Claude Mythos 5.2 is the same underlying model without certain cyber and biology safeguards. Access is limited to vetted organizations in Anthropic's Cyber Verification Program and Life Sciences Verification Program.
• Claude Enterprise customers can use Claude Security, which runs on Mythos 5.2 and does not require membership in either program, according to Anthropic.

Fable 5.2 vs Mythos 5.2

Source for this table: Anthropic's announcement.

What changed, according to Anthropic

Speed and fast mode

Anthropic says outputs are up to 30% faster than Fable 5.1 and Mythos 5.1. An optional fast mode speeds up token generation by up to 2.5 times. It is available in Claude Code, on the Claude Platform, on Claude Models on Google Cloud, and on Anthropic-hosted Microsoft Foundry.

API pricing

Anthropic lists standard prices of $8 per million input tokens and $40 per million output tokens for both models, and $16 and $80 in fast mode. Anthropic calls this a lower price. VentureBeat reports that it is a 20% cut from the $10 and $50 charged for the 5.1 models.

Coding and reasoning benchmarks

Anthropic's published comparison reports these results for Fable 5.2, set against Fable 5.1:

• FrontierCode 2.0: 49.1% at Xhigh effort, compared with 32.4%.
• FrontierSWE v2: 69.9%, compared with 56.3%.
• Humanity's Last Exam (with tools): 74.2%, compared with 65.6%.

Anthropic's own table also shows competitors ahead on some lines. On AutomationBench, for example, it lists Gemini 4 Carbon at 53.9% against 51.4% for Fable 5.2. Opus 5.5 is ahead on one Elo ranking row (1,800 against 1,785).

Cybersecurity and Mythos 5.2

On a CVE Rediscovery test (an October 2026 run covering 1,829 vulnerabilities), Anthropic reports that Mythos 5.2 rediscovered 89.6% of known vulnerabilities. It reports 78.3% for Mythos 5.1 and 81.3% for Opus 5.5 with its safeguards disabled.

Safeguards

Anthropic says Fable 5.2's cyber and biology safeguards are as robust as Fable 5.1's but block legitimate requests less often: about 55% fewer false blocks overall in early testing. Blocked requests fall back to Opus 4.8 for cyber topics and Opus 5 for biology topics.

Science and engineering examples

Anthropic describes several experiments, all from its own testing:

• Hardware: Fable 5.2 designed three electronic devices (a home voice-assistant hub, a USB speakerphone and a handheld virtual pet). Anthropic says each board was ready for manufacturing in under a day and that all three worked without electrical modifications. The work was done with Diode.
• Robotics: On RoboDojo Sim, Anthropic reports 64.9% for Fable 5.2, compared with 27.6% for Fable 5.1.
• Theoretical physics: Anthropic says Fable 5.2 solved two open problems about black holes.

What to keep in mind

• These are Anthropic's results. The announcement and system card are the source for the benchmarks, speed claims and experiments above. We have not independently reproduced them.
• VentureBeat notes some caveats. The customer testimonials in the launch were selected by Anthropic. For FrontierCode, Cognition reported no uncertainty intervals. Labs used different tasks in the CyScenarioBench comparison, so it is a looser comparison.
• Opus 5.5 is still recommended for most work. VentureBeat reports that Anthropic recommends Opus 5.5 for most coding and knowledge work.
• The announcement page was edited after publication. Anthropic amended one sentence and added a footnote on November 11. We have not relied on that section.
• Not stated in the announcement: a context window size, rate limits, or regional availability.

Why it matters

This section is our analysis, not a reported fact.

The release shows two trends. Providers are competing on speed and cost as well as on capability, and the price cut and fast mode both point that way. They are also splitting access by risk, with a broadly available model and a restricted one built on the same base. For developers, a benchmark score is a starting point. Latency, cost per task, reliability, access and safety restrictions on your own workload should decide the choice.

Frequently asked questions

Is Claude Fable 5.2 available to everyone?

Anthropic says Fable 5.2 is available on all platforms, including AWS, Google Cloud and Microsoft Azure. Plan and platform details can change, so check Anthropic's page.

What is Claude Mythos 5.2?

It is the same underlying model as Fable 5.2 without certain cyber and biology safeguards, limited to vetted organizations in Anthropic's verification programs.

Is Fable 5.2 faster than Fable 5.1?

Anthropic reports outputs up to 30% faster than the 5.1 models, and up to 2.5 times faster token generation in fast mode.

Can developers use Fable 5.2 through an API?

Yes. Anthropic lists the API model string claude-fable-5-2 for Fable 5.2.

Product details and availability may change after publication. Check the official announcement for the latest information.