Anthropic Releases Claude Fable 5.2 and Claude Mythos 5.2

Anthropic Releases Claude Fable 5.2: System Prototype Developed in Eight Days, Output Speed Increased by 30% Phoenix Tech Download the app for exclusive early access November 10, 2026 06:48:07 From Beijing Updated November 10, 2026 07:23:46

Anthropic Releases Two New Models

Phoenix Tech News, Beijing Time, November 10 (Reuters) — Anthropic on Monday released two new AI models, Claude Fable 5.2 and Claude Mythos 5.2, hoping to consolidate its position in the fast-growing AI market ahead of its planned IPO.

Of the two new models, Claude Fable 5.2 is mainly aimed at coding, reasoning, and AI agent tasks, while Claude Mythos 5.2 is aimed at approved cybersecurity and life sciences research applications.

Anthropic said in its official blog that the two models are designed to tackle the toughest technical challenges. Claude Fable 5.2 and Claude Mythos 5.2 show improvements in areas including coding, reasoning, and vision. In tests before release, they helped Anthropic's engineers complete software migration projects that were previously difficult to achieve, design three working electronic devices from scratch, solve two unresolved problems in black hole physics, make major progress in robotics evaluations, and complete a formal proof of Fermat's Last Theorem within 30 hours.

Claude Fable and Claude Mythos are based on the same underlying model. Fable is equipped with strict cybersecurity and biosecurity safeguards to prevent misuse, so it can be opened to all users. Organizations engaged in beneficial applications in cybersecurity or biomedical research can apply for access to Mythos through Anthropic's recently expanded life sciences and cyber verification programs.

Testers found that, compared with the previous Fable and Mythos models, Fable 5.2 and Mythos 5.2 communicate more clearly with users. Their output speed is up to 30% faster than Claude Fable 5.1 and Claude Mythos 5.1. The new models are also cheaper: $8 per million input tokens and $40 per million output tokens.

Fable 5.2 Leads on Multiple Benchmark Metrics

Benchmark tests show that Claude Fable 5.2 and Mythos 5.2 outperform Opus 5.5 on almost every capability Anthropic evaluated.

At medium effort (the default setting for the Claude app and Claude Code), Fable 5.2 can complete challenging tasks using fewer tokens than Fable 5.1. In early tests of the FrontierCode 2.0 evaluation, Fable 5.2 at high effort scored roughly on par with Opus 5.5 at maximum effort, while costing about half as much per task. At ultra-high effort, Fable 5.2 also scored higher than all released models participating in the test.

Today, more than 90% of the changes merged into Anthropic's core codebase are written by Claude (under human guidance). Anthropic has used Fable 5.2 to fix stubborn software bugs in minutes, optimize the performance of all services, and complete high-value software projects that were previously too slow or too costly to attempt. A group of agents running an early version of Fable 5.2 migrated a critical production service from Python to Rust, and it is now more efficient and reliable, with startup about 26 times faster. Anthropic found that Fable 5.2 demonstrates strong engineering judgment, producing high-quality results even when faced with ambiguous requirements.

As an experiment, an Anthropic engineer asked Fable 5.2 to design an operating system prototype. Starting from an empty folder, the model coordinated more than 1,000 sub-agents, writing, testing, and reviewing code around the clock for eight consecutive days. In the end, it developed an operating system that can boot and run on a laptop, with its own kernel for running the machine, drivers for disk, keyboard, and network, persistent file storage, windows, a login screen, an installer, and built-in Claude features.

Claude Mythos 5.2 is Anthropic's most capable model to date in cybersecurity engineering. Compared with previous models, it can find more security vulnerabilities and is better at reproducing and fixing them. Note that Claude Fable 5.2's safeguards block most cybersecurity-related tasks.

A cybersecurity company found that one-sixth of the vulnerabilities Claude Mythos 5.2 confirmed finding were missed by all the other frontier models they tested. In Anthropic's internal CVE Rediscovery¹ evaluation, Claude Mythos 5.2 found 10.2% more vulnerabilities than Claude Opus 5.5 and 14.4% more than Claude Mythos 5.1.

Claude Mythos 5.2 also shows clear improvements in judgment and assessment accuracy. One tester found that previous models exaggerated the severity of the problem in about 1 out of every 20 vulnerability findings, while Claude Mythos 5.2 showed no exaggeration at all.

Fable 5.2 and Mythos 5.2 also offer an optional fast mode that can increase token generation speed by up to 2.5 times. Fast mode is priced at $16 per million input tokens and $80 per million output tokens. (Author/Xiao Yu)

For more first-hand news, please download the Phoenix News app and subscribe to Phoenix Tech. For in-depth reporting, search "Phoenix Tech" on WeChat.