I Tested Fable 5.2 vs Opus 5.5 on 5 Real Use Cases

I Tested Fable 5.2 vs Opus 5.5 on 5 Real Use Cases

More

Summary

Peter Yang compares Claude Fable 5.2 against Opus 5.5, previously Anthropic’s best model, across five real-world use cases. Each model gets the same prompt and two attempts: an initial build, then a self-review and one revision.

The tests include a 20-second 3D marketing video of Kyoto’s Fushimi Inari Shrine made entirely in code, a Mario Kart-style 3D cart racer built with Three.js and Blender-made models, a conversion-focused homepage redesign, and a writing test in which each model produces five 600-word essays on the most underrated human skill in the AI era. Yang counts AI-slop words across the essays and also probes how well each model holds its position when he pushes back on its advice.

Fable’s video and racing game come out more polished, though it takes longer: about 62 minutes for the game versus 34 for Opus. Fable gets a slight edge on the website redesign, although Yang says both models overload pages with copy. He also reports one test where Opus still comes out ahead. The video is a practical look at where the upgrade is worth paying for.


📺 Source: Peter Yang · Published November 09, 2026
🏷️ Format: Comparison

1 Item

Channels