Anthropic unveils ‘Fable 5.2,’ which leads 1,000 AIs as a team [Global AI Briefing]
- Input
- 2026-11-10 08:32:47
- Updated
- 2026-11-10 08:32:47
This marks a shift beyond AI handling individual tasks at people’s direction to a stage where it can independently break down complex projects and collaborate.
On the 9th (local time), Anthropic unveiled its latest AI models, Claude Fable 5.2 and Claude Mythos 5.2. Both are specialized in demanding technical tasks such as coding, reasoning and scientific research.

Fable 5.2 is available to general users and developers and has safeguards that block dangerous tasks in cybersecurity and biology. Mythos 5.2 is a model with some of these restrictions lifted and is being made available on a limited basis to vetted cybersecurity and life sciences research institutions.
Anthropic particularly highlighted Fable 5.2’s collaborative capabilities.
According to Anthropic, when an Anthropic engineer instructed Fable 5.2 to develop an operating system prototype, it created at least 1,000 subordinate AI agents to carry out the work. Anthropic said Fable 5.2 divided up the roles among AIs, including ones that wrote the operating system’s core code, implemented storage, keyboard and network functions, and found and fixed errors.
After eight days of work, the AI team completed an operating system that actually boots on a laptop. Anthropic said it also included a display, login and software installation features.
Anthropic said Fable 5.2 scored 49.1% on a test assessing the likelihood that open-source software developers would adopt its code. That was up from 32.4% for the previous version, Claude Fable 5.1.
The unveiling of the new model is seen as a move by Anthropic, which is preparing for an IPO as early as this month, to highlight its technological capabilities and show investors its potential to attract corporate customers. The strategy is to expand its market by improving performance and lowering usage fees.
[email protected] Lee Gu-sun Reporter