Claude Fable 5.2 and Mythos 5.2 have arrived

Anthropic announced Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026. The company says the models improve in coding, reasoning and vision; The standard price was announced as 8 dollars for million input tokens and 40 dollars for million output tokens.

According to @claudeai post, the models are designed to solve challenging technical tasks and aim for stronger performance in collaborative projects. Additionally, Anthropic's statement stated that Fable 5.2 is open to everyone, while Mythos 5.2 is for verified organizations.

What's the difference between Claude Fable 5.2 and Mythos 5.2?

According to Anthropic, the two models share the same basic structure; However, usage policies differ. For example, Fable 5.2 is being made available to the general public with stronger protections against cybersecurity and biology abuse, while Mythos 5.2 is being made available to verified organizations through the Life Sciences and Cyber ​​Verification Programs.

The company also said that output is up to 30 percent faster than Fable and Mythos 5.1, and token generation is accelerated up to 2.5 times with the optional fast mode. Therefore, the fast mode price is 16 dollars for million input tokens and 80 dollars for million output tokens.

What about the price and availability of Claude Fable 5.2?

Anthropic recommends starting with Claude Opus 5.5 for day-to-day coding and computing, and moving to Fable or Mythos as the scope grows. On the other hand, Sonnet 5.5 and Haiku 5.5 are recommended for simple works. Fable 5.2 can also be used in the Advisor or Orchestrator role. The API ID is claude-fable-5-2 and the model is available today on all platforms including AWS, Google Cloud and Microsoft Azure.

Türkiye price or local access date has not been announced. Additionally, Fable usage quota is also offered as a promotion for paid plans this week.

What performance examples did Claude Fable 5.2 share?

According to Anthropic's own table, Fable 5.2 scored 49.1 percent in the FrontierCode 2.0 Xhigh test; 69.9 percent in FrontierSWE v2; Humanity's Last Exam gave a 74.2 percent result in instrumental measurement. Additionally, the Terminal-Bench-Science 0.1 score for Mythos 5.2 was shared as 75.3 percent, and the CyScenarioBench score was shared as 76.3 percent. However, these numbers are the company's own tests; There is no independent third-party verification for this news.

On the security side, Anthropic said that Fable 5.2's false positives were reduced by approximately 55 percent in external evaluations conducted with the UK AI Security Institute, Frontier Design and METR. For example, this rate was announced as approximately 60 percent in the cyber field and approximately 40 percent in the biology field. Bridgewater, Salesforce and Stripe are included in the Enterprise Frontier Safeguards private beta program.

ModelPriceAccessNote Claude Fable 5.2$8 / $40General useclaude-fable-5-2 Claude Mythos 5.2$8 / $40Verified organizationsVerification programs Fast mode$16 / $80OptionalUp to 2.5x speed

You can also review our news about the previous version, Claude Fable 5.1. For another development on the artificial intelligence side, you can also read the GPT-6 Cyber ​​news. Here is the Claude Fable 5 news for Anthropic's previous announcement.