NVIDIA Introduces Nemotron for Science to Help Researchers Build Scientific Discovery Agents
- Published
- Source
- NvidiaNews
NVIDIA introduced Nemotron for Science, a package combining scientific applications of its Nemotron 3.5 Super VL open model, an open chemistry dataset and post-training recipes. The offering is intended to help researchers build specialized agents that reason through scientific problems and use research tools. Its LitmusBench v0.2 dataset contains nearly 50,000 questions for training and evaluating chemistry reasoning. NVIDIA said life sciences companies Lila Sciences, Terray Therapeutics and Therna are using or customizing the model, while Argonne and Jülich Supercomputing Centre are integrating it into scientific and drug-discovery agents.
The examples illustrate how domain-specific post-training can support research workflows, from extracting chemical information and planning experiments to designing therapeutics and laboratory protocols. Terray Therapeutics reported that its post-trained model surpassed previously published results on two benchmarks and recovered about 76% of annotated reactions on the RxnScribe test under its matching criteria. Lila Sciences reported 61% accuracy for making quantum dots after laboratory-feedback training, compared with 54% for GPT-6 Astra, at 10–20 times lower inference cost. Therna reported nearly doubled accuracy in predicting protein-expression ranges for agentic mRNA therapeutic design.
Argonne is integrating Nemotron into its StructBioReasoner multi-agent system to design and evaluate protein-based therapeutic candidates for cancer and other diseases. The source does not disclose a completion date for this integration. Further reported performance results include scientists preferring Therna’s model explanations over the base model’s in 96% of comparisons.
Original report
NVIDIA Introduces Nemotron for Science to Accelerate Discoveries With Agents
NvidiaNews · NVIDIAThis briefing is an independently written summary based on publicly available reporting and is provided for industry information and news discovery. The original report and source publication are credited and linked where applicable. RobotToday does not claim ownership of third-party source material.
Rights concerns? Contact [email protected] with the relevant URL and details. We will review the matter and take appropriate action where warranted.
More related news
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
-
12
RobotToday Weekly November 9–13, 2026 RobotToday · 3 days ago
Related Suppliers
NVIDIA Robotics
NVIDIA robotics division delivering Isaac simulation and training tools, Jetson edge compute modules, and Omniverse digital twin frameworks for physical AI.
Deci AI (Acquired by NVIDIA)
Israeli deep learning optimization startup acquired by NVIDIA in May 2024 for ~$300M; known for AutoNAC neural architecture search technology.
Allen Institute for Artificial Intelligence
Seattle non-profit AI institute founded by Paul Allen; releases fully open robot models (MolmoAct, MolmoBot) and simulators (AI2-THOR, MolmoSpaces).
Aether AI
Seed-stage San Diego startup building causal world models that predict action outcomes, with robotics as its first target market.
Scentience
US startup designing neuromorphic olfactory sensors, edge AI models, and developer APIs to give robots a sense of smell for navigation, security, and.
Alfa Chemistry Materials
US chemical and materials supplier offering a broad catalog of 3D printing materials, including engineering plastics, resins, ceramics, and metals for R&D.
Arcee AI
U.S. AI model lab building efficient open-weight small language and foundation models that run on edge, on-prem or cloud.
Anyscale
Anyscale is a San Francisco AI infrastructure firm behind Ray, the open-source engine for large-scale model training; Nscale agreed to buy it for $1.65B.
Medra
San Francisco startup building Physical AI lab robots that autonomously run drug discovery experiments end-to-end in an autonomous 38,000 sq ft lab facility.
Iktos
Paris-based Iktos pairs generative AI drug design with Iktos Robotics, an autonomous lab running about 100 robotic reactions per day on Chemspeed hardware.
LS2N - Laboratoire des Sciences du Numérique de Nantes
LS2N is a Nantes digital-science lab of 534 members whose ARMEN, ReV and RoMas teams research autonomous, soft and manufacturing robots.
ATS Corporation
Canadian factory automation company (TSX/NYSE: ATS, est. 1978) building custom assembly and test systems for life sciences, EV, and semiconductor sectors.
Robotics needs a service framework.
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
Join the RobotToday community on LinkedIn
Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.