Deep Principle Releases Hydrogen-1.0, Open-Weight AI Model for Scientific Research

Deep Principle says open-weight AI for science enables researchers to inspect and adapt models, contrasting with proprietary approaches. Hydrogen-1.0 builds on GLM-5.3 and was trained on lab records and discovery trajectories. The company reports that it scored 38.3% on FrontierScience-Research versus GPT-6 Astra's 33.3%, and 93.0% on GPQA-Diamonds. Demonstrations include simulation debugging and experimental protocol planning.