Confirmed · Nov 4, 2026, 09:00 UTC

Deep Principle releases open-weight Hydrogen-1.0 science model built on GLM-5.3

The company says Hydrogen-1.0 scores 38.3% on FrontierScience Research versus 33.3% for GPT-6 Astra at max thinking, and 92.9% on GPQA Diamond versus Kimi K3's 93.5%. Cost figures are estimates under the company's own evaluation setup, and the cases are selected walkthroughs.

• LabDeep Principle
• ModelHydrogen-1.0
• ModelGLM-5.3
• ModelGPT-6 Astra
• ModelKimi K3
• Topicrelease
• Topicbenchmark
• Topicopen-weights
• Topicresearch