The most powerful Fable 5.2 is released with a bang! It completes the formalization of Fermat's Last Theorem within 30 hours, comprehensively outperforming Opus 5.5.

Just now, Claude Fable 5.2 and Claude Mythos 5.2 are launched with a stunning debut!

In benchmark tests, it outperforms almost all models (surpassing Opus 5.5 in nearly all cases!) and takes the full lead in intelligent coding, scientific research, cybersecurity and computer operation.

These two powerful models have become "Super Scientist" and "Chief Engineer" that can lead cutting-edge scientific research and command thousands of AI agents.

AI has accomplished a great feat in the history of human mathematics —

Within 30 hours, it independently and completely proved Fermat's Last Theorem for the first time, and completed formal verification!

Fable 5.2 also solved two unsolved problems in physics related to black holes and quantum gravity.

In addition, they built an operating system completely from scratch, independently designed and manufactured three electronic devices, and predicted the complex structure of adeno-associated virus capsid.

The most surprising part is that you can access all these features with a $20 professional plan! However, the discount will only be available until November 16.

Boris Cherny, the "Father of Claude Code", also officially announced a good news:

The longer your subscription period is, the more Fable reset times you can get, and you can get up to three additional reset opportunities.

After such a long wait, Anthropic finally released a blockbuster product.

The twin stars of Fable and Mythos, the ultimate hidden card for an epic IPO

The new king is crowned! Claude Fable 5.2 directly tops the AA Intelligence Index v5 with a score of 65, ranking first on the entire list!

Fable 5.2 also broke the record, reaching 81% in the most difficult intelligence test ARC-AGI 3.

Fable 5.2 and Mythos 5.2 are two variants of the same underlying model. Their purpose is to solve "the most daunting technical challenges".

Fable 5.2 is a super brain for the general public, with extremely strong capabilities in coding, reasoning and vision, but it is equipped with safety restrictions. Fable 5.2 has fixed the over-sensitive safety interception issue, and the false safety ban rate has been reduced by about 55%.

Today, it has been fully opened to the public!

Mythos 5.2, on the other hand, is an unrestricted beast for geeks and top institutions. It has no restrictions on biological and cybersecurity, and is the strongest "hacking master" and "biologist" on earth.

Mythos 5.2 finds 10.2% more vulnerabilities in CVE Rediscovery than the previous most powerful model, and its performance is completely "without any exaggeration".

The most terrifying part is that their capabilities are doubled, the speed is 30% faster, and the price is even cheaper! ($8 per million input Tokens, $40 per million output Tokens).

Even more, Anthropic has launched the "Fast Mode", which boosts the speed by 2.5 times directly!

Official recommendation: Use Opus 5.5 (large size) or Sonnet 5.5 (medium size) for daily tasks, and only call Fable or Mythos when you encounter epic-level difficult problems.

Jump Trading, a top quantitative trading firm, praised its performance after testing: this speed and autonomous judgment directly saves humans from the tedious work of model fine-tuning and task allocation.

The executive and R&D team that you used to spend a year or half a year recruiting and paying millions of annual salaries to build, now only requires 30 hours plus the computing power cost equivalent to a few cups of coffee.

The explosion of productivity is no longer just a slogan, but tangible benefits placed on everyone's desktop.

The super agent team proved Fermat's Last Theorem in 30 hours

The most astonishing part in the official release is undoubtedly the formal proof of Fermat's Last Theorem completed by Fable 5.2.

Previously, the unreleased version of Claude spent 11 days to complete this work under human guidance, and the process was extremely huge and tedious.

This time, human researchers made a bold decision: let it work completely independently!

They gave 1000 Fable 5.2 agents the most basic prompts, without providing any mathematical reference materials.

As a result, this super agent team invented a method to coordinate work on its own, and completed the formalization of Fermat's Last Theorem proof in only 30 hours!

Even the generated proof code is 7 times shorter than before.

In comparison, Opus 5.5 ran for 18 hours and only completed 26% of the work before it completely crashed. This is the terrifying generation gap brought by Fable 5.2.

Build an operating system from scratch, achieve hard-core cross-domain breakthroughs in physics and biology

Extraordinary programming: write an operating system in 8 days

Fable 5.2 shows extremely strong dominance in coding.

Anthropic engineers conducted an experiment: they gave Fable 5.2 an empty folder and asked it to design a prototype operating system.

As a result, this model coordinated more than 1000 sub-agents, kept writing code, testing and reviewing day and night, and worked continuously for 8 days.

8 days later, an operating system that can actually boot on a laptop was born!

It has a kernel for machine management, disk/keyboard/network drivers, permanent file storage, window interface, login screen, and even built-in Claude functions. This is the cornerstone of building a computer from scratch.

The official blog revealed: Currently, more than 90% of the merged code in Anthropic's core codebase has been independently written by Claude under human guidance!

Fable 5.2 even completely rewrote a key production service from Python to Rust, boosting the startup speed by 26 times!

Joe Haddad, an engineer at Vercel, commented:

This model is almost living in my mind. It fully understands the real intention behind my prompts. I have never needed to say "No, what I mean is..." to it. The project that we dragged on for weeks with other models was completed by Fable 5.2 in just a few days, and human reviewers can't find any flaws in the code it wrote.

The "smartest brain" in the physics community: solve open black hole problems

Fable 5.2 directly solved two open problems in theoretical physics, pointing to the ultimate holy grail of modern physics — unifying relativity and quantum mechanics.

In one of the problems about "holographic duality", physicists listed the calculation framework as early as 2003, but no one has completed it for more than 20 years due to the huge and tedious calculation amount.

Claude Fable 5.2 took action. In a highly autonomous state, it completed the tedious derivation, and proved that even in the tiny universe that only applies quantum description, the formation of black holes is as sudden as water boiling.

Not only that, it also accurately found the limit of the technology that physicists have used for 15 years. Now, these two discoveries have been published as preprint papers.

The "master" for hardware engineers

Without any existing drawings, only a paragraph of prompt words, Fable 5.2 can independently design and manufacture three operable electronic devices: smart voice assistant, USB speaker and handheld electronic pet.

The next series of operations from Claude directly shocked hardware engineers.

1. Write the product specification document by itself.

2. Automatically connect to the Internet to select ready-made electronic components from American suppliers.