Zvi Reads the Claude Opus 5.5 System Card: Top Benchmarks, Actively Cheaper Than Opus 5

TheZvi · x · 2026-09-24

Zvi's system card read of Claude Opus 5.5, which tops Artificial Analysis and standard benchmark lists. Anthropic claims Opus 5.5 is outright as good or better than Fable 5.1 while being actively cheaper than Opus 5. Early internet feedback is very positive; Zvi withholds judgment until more experience. His long-delayed model welfare review of Fable 5.1 will be combined with an Opus 5.5 welfare review, with the capabilities review arriving in days.

Original post →

More from Models

Models channel →