Subscribe

Nemotron for Science

NVIDIA announced Nemotron for Science at SC26, a toolkit that pairs the Nemotron 3.5 Super VL open model with an open chemistry dataset and post-training recipes for science agents.

Nemotron for Science combines the Nemotron 3.5 Super VL open model with a new open chemistry dataset and post-training recipes for life sciences and chemistry. Researchers can run it locally on a DGX Spark, a workstation or their own supercomputer. NVIDIA says an agent built with it at Argonne National Laboratory planned and tested hypotheses from one prompt and identified two new binding sites.

Date
Monday 16 November 2026
Lab
NVIDIA
Kind
open-weights
Access
open weights

Figures

MeasureValueMeasured by
GROMACS test cases fixed by KTH agent81%
one overnight run on a single DGX Spark
company
WRF routine speedup from Nsight AI Porting Agentmore than 5x
about 150 million tokens used; a separate NVIDIA agent, not part of Nemotron for Science
company

Announced in Ian Buck's SC26 address; license terms, download location and exact release date were not stated on the page read, and 'open weights' is inferred from the description as an open model.

Sources

  1. blogs.nvidia.com/blog/ian-buck-sc26/

This record was partly confirmed: some claims could not be checked on 17 November 2026. How we check

Read the daily brief for 16 November 2026