Opus 5.5 reportedly an RL run on the Opus 5 pretrain, trained in ~2 months

haider1 · x · 2026-09-28

Leaker haider claims Anthropic's Opus 5.5 is a reinforcement learning run on top of the Opus 5 pretrain, taking about 2 months. Reportedly even Anthropic was surprised by the results: naming shifted from 5.1 to 5.2 and finally 5.5 as it kept getting stronger than expected. Unconfirmed rumor.

Original post →

More from Models

Models channel →