DaisyRose-27B: Pushing the Frontier of Post-Training EfficiencyNew
Introducing DaisyRose-27B, RadixArk’s first model, trained with Miles on 1,200 curated tasks in 18 days on 144 H200 GPUs.
Engineering, open source, and news from RadixArk.
Introducing DaisyRose-27B, RadixArk’s first model, trained with Miles on 1,200 curated tasks in 18 days on 144 H200 GPUs.
RadixArk joins the SGLang Consortium as a founding member. Read the announcement on the LMSYS blog.
Miles v0.1 is an open-source RL framework for LLMs and multimodal models, combining fast multi-turn SGLang rollouts, fully asynchronous RL, and Token In Token Out.
RadixArk announces $100 million in Seed funding at a $400 million post-money valuation to build open infrastructure for frontier AI, powered by SGLang and Miles.