Hydrogen-1.0 MIT Open Weights: Gated BF16, NVFP4 262K Context Cliff, and Vendor FSR vs GPT-6 Astra
Attributed compile + one small offline run on my box Primary source: Introducing Hydrogen-1.0 by Deep Principle , published 2026-11-02 . Weights: deep-principle/Hydrogen-1.0 (BF16, gated: auto, MIT, base zai-org/GLM-5.3 ) · Hydrogen-1.0-NVFP4 (ungated) · Hydrogen-1.0-MXFP4 (ungated). Local USAGE / NVFP4 card copies: /workspace/reputation/runs/32-hydrogen-1.0/ . Secondary compile: forkalt —…
Deep Principle unveiled Hydrogen-1.0 on November 2, 2026. It is an open-weight scientific model designed for chemistry, materials science, and life sciences, created by fine-tuning the zai-org/GLM-5.3 model. Hydrogen-1.0 comes in three variants: BF16 gated, ungated NVFP4, and ungated MXFP4. The model offers vendor-reported benchmarks, stating it outperforms GPT-6-Astra on FrontierScience-Research (FSR) and improves over GLM-5.3 on GPQA-Diamonds.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
