Skip to content
Sıfırıncı Dakika
Latest

NVIDIA introduces Nemotron for Science to power scientific discovery agents

AILaunchUpdated: 1 min read
NVIDIA Newsroom

In brief

NVIDIA unveiled Nemotron for Science, a package built on its open Nemotron 3.5 Super VL model. It bundles LitmusBench v0.2, an open chemistry dataset of nearly 50,000 questions, along with post-training recipes and infrastructure from DGX Spark to data centers. Research centers such as Argonne and Jülich are using the model in scientific discovery agents.

  • Built on the open Nemotron 3.5 Super VL model
  • LitmusBench v0.2, an open chemistry dataset with nearly 50,000 questions
  • Post-training recipes for domain-specific adaptation
  • Infrastructure from DGX Spark desktops to data centers
  • In use at institutions including Argonne and Jülich

A science-focused agent package

NVIDIA introduced Nemotron for Science, built on its open Nemotron 3.5 Super VL model. The goal is to help researchers build specialized agents for their own fields. The package combines the model with evaluation sets, an open chemistry dataset and post-training recipes.

LitmusBench v0.2 and chemistry data

At the center is LitmusBench v0.2, an open dataset of nearly 50,000 questions. Drawing on molecular properties and experimental drug discovery data, it is designed both to evaluate and to improve models' chemistry reasoning. Researchers can use it to boost a model's ability to predict how molecules behave, including solubility and how the body processes them.

Post-training and notable results

Researchers can post-train Nemotron models with domain data to specialize them. According to NVIDIA's examples, Terray Therapeutics trained the model to extract reactants, products and reaction conditions from chemistry diagrams, recovering about 76% of annotated reactions on the RxnScribe test. Lila Sciences raised accuracy for making quantum dots to 61% for its automated labs; the company says this beats GPT-6 Astra's 54% at 10-20x lower inference cost.

Who is using it

Institutions including the U.S. Department of Energy's Argonne National Laboratory and the Jülich Supercomputing Centre are integrating the model into drug and scientific discovery agents. Lila Sciences, Terray Therapeutics and Therna are customizing it for science applications.

Infrastructure

NVIDIA points to DGX Spark and data center infrastructure to take Nemotron 3.5 Super VL from researchers' desktops to supercomputing centers, letting organizations run secure scientific workflows on infrastructure they control.

Why it matters

AI agents are spreading fast in scientific discovery; this matters to readers because open models and datasets let researchers build domain-specific agents of their own.

Sources

Related stories