# Khazix tests Claude Fable 5.2 on video, 3D web, and design tasks

> Original title: 实测Claude Fable 5.2 - 审美与细节的绝对统治。

- Source: Khazix
- Published: 2026-11-09T23:53:55.000Z
- Bad Influence: https://badinfluence.news/items/cyaamrkal5hm0owea46uhat9o
- Original: https://x.com/Khazix0918/status/2119941585538572390

## Summary

Khazix says Anthropic released Claude Fable 5.2 at 3 a.m. with a 1M-token context window, 128K output, and API pricing cut 20% to $8 input and $40 output per million tokens. In his own tests, Fable 5.2 generated a ChatGPT four-year tribute video, a DeepSeek explainer video, and a Three.js Apple Shanghai Jingan store scene, which he says drew on about two and a half hours of work each. He says the tests consumed roughly 24% of his Ultra subscription quota and calls the model expensive.

## Post

https://x.com/i/article/2119938260130832384
