MiniMax's Open Weight Model Strategy and Inference Stack

About

Olive Song, who leads reinforcement learning at MiniMax, discusses the company's philosophy of releasing open weight models and details the engineering behind them, including RL training for agentic coding and computer-use tasks against environments like OS World, custom GPU kernel work benchmarked via a parallel kernel bench, and day-zero inference readiness. She also covers a multimodal training pitfall where text and vision performance collapse unless both modalities are trained jointly, and reflects on scaling toward longer-horizon agent tasks like a twelve-hour run.