RadixArk trains DaisyRose-27B in 18 days on 144 H200s, doubling Terminal-Bench score

BanghuaZ · x · 2026-11-19

RadixArk released its first open-weight model, DaisyRose-27B: RL training on top of Qwen3.8-27B with 1,200 curated long-horizon agent tasks, using 144 H200s for just 18 days.

A reference case of a small team massively boosting a small model's agent capability with short-cycle RL.

Original post →

More from coding & agent

coding & agent channel →