Opus 5.5 field results: 200k-line codebase fixed in under 3 hours at 1/5 the cost of GPT-6 Astra

vista8 · x · 2026-09-23

A detailed roundup of Claude Opus 5.5 benchmarks and field tests:

Benchmarks

Coding efficiency

Knowledge-work accuracy

Cost & misc

Related event: Claude Opus 5.5 Launches to Top Rankings with Lower Price and Faster Speed(28 posts)→

Original post →

More from Models

Models channel →