# Anthropic releases Claude Fable 5.2 and Claude Mythos 5.2 for difficult technical work

> Original title: Anthropic 发布 Claude Fable 5.2 和 Claude Mythos 5.2，面向最难啃的技术工作，并在编程、推理和视觉上有提升。btw... 终于还是没等到 GPT 6.2 ...

- Source: meng shao
- Published: 2026-11-10T00:20:10.000Z
- Bad Influence: https://badinfluence.news/items/ay9hoo2htmgewyffre7najtr2
- Original: https://x.com/shao__meng/status/2119948191793861080

## Summary

Anthropic has released Claude Fable 5.2 and Claude Mythos 5.2, which share one underlying model, according to meng shao's post. Fable has stronger cybersecurity and biosecurity safeguards and is available on paid plans and major cloud platforms. Mythos requires approval through a life sciences or cybersecurity verification program and is aimed at beneficial biomedical or cybersecurity research. The post says Fable 5.2 scored about 45.3% on FrontierCode 2.0 at High effort, close to Opus 5.5's 45.5%, and 49.1% at Xhigh effort.

## Post

Anthropic 发布 Claude Fable 5.2 和 Claude Mythos 5.2，面向最难啃的技术工作，并在编程、推理和视觉上有提升。btw... 终于还是没等到 GPT 6.2 ...

Claude Fable 5.2 和 Claude Mythos 5.2 是同一套底层模型。Fable 带有更强的网络安全和生物安全护栏，已在付费方案和主流云平台上线；Mythos 需通过生命科学或网络安全验证项目申请，面向有益的生物医学或网络安全研究。日常多数编程和知识工作，官方仍建议先用 Opus 5.5，复杂、开放的长程任务再换到这一档。

https://www.anthropic.com/claude-fable-and-mythos-5-2

对齐上，多数指标不低于近期模型，并由英国 AI 安全研究所、Frontier Design、METR 等做了发布前外部评测。能力上，几乎所有评测维度都高于 Opus 5.5。在 FrontierCode 2.0 上，Fable 5.2 开到 High 档约得 45.3%，与 Opus 5.5 的最高分 45.5% 大致持平，单任务成本约一半；开到 Xhigh 档得 49.1%，是已测已发布模型里最高的。应用默认是 Medium 档。

演示性结果：约 1000 个 Fable 5.2 智能体用 30 小时完成费马大定理形式化；同样设置下，Opus 5.5 约 18 小时完成 26%，30 小时后几乎不再推进。Mythos 在尝试的 34 个基因疗法衣壳中，准确建模了 28 个。仿真双臂基准 RoboDojo 上，内部最高档跑分为 64.9%，高于 Opus 5.5 的 39.4% 和 Fable 5.1 的 27.6%。

## AI translation

Anthropic has released Claude Fable 5.2 and Claude Mythos 5.2, aimed at the hardest technical work, with gains in coding, reasoning, and vision. btw... still didn't get GPT 6.2 ...

Claude Fable 5.2 and Claude Mythos 5.2 are the same underlying model. Fable comes with stronger cybersecurity and biosecurity safeguards and is available on paid plans and major cloud platforms; Mythos requires applying through a life sciences or cybersecurity verification program and is aimed at beneficial biomedical or cybersecurity research. For most everyday coding and knowledge work, the official recommendation is still to start with Opus 5.5, and switch to this tier for complex, open-ended long-horizon tasks.

https://www.anthropic.com/claude-fable-and-mythos-5-2

On alignment, most metrics are no lower than recent models, and external pre-release evaluations were conducted by the UK AI Security Institute, Frontier Design, METR, and others. On capability, nearly every evaluation dimension is above Opus 5.5. On FrontierCode 2.0, Fable 5.2 at High effort scores about 45.3%, roughly even with Opus 5.5's top score of 45.5%, at about half the per-task cost; at Xhigh effort it scores 49.1%, the highest among tested and released models. The app default is Medium effort.

Demonstration results: about 1,000 Fable 5.2 agents formalized Fermat's Last Theorem in 30 hours; under the same setup, Opus 5.5 reached 26% in about 18 hours and barely advanced after 30 hours. Mythos accurately modeled 28 of 34 attempted gene therapy capsids. On the simulated dual-arm benchmark RoboDojo, the internal highest tier scored 64.9%, above Opus 5.5's 39.4% and Fable 5.1's 27.6%.

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWuRo1daoAA-t7P.jpg&mode=thumb&exp=1794528000&sig=89b3ea9f28c34c8a)

## Quoted post @claudeai

> Introducing Claude Fable 5.2 and Claude Mythos 5.2, designed to tackle the most daunting technical challenges.
> 
> They’re capable collaborators on ambitious projects in many fields, and show improvements in areas like coding, reasoning, and vision.

https://x.com/claudeai/status/2119867114450268322