NVIDIA releases Nemotron 3.5 Super VL open science model and datasets
11/17/2026 — 11/17, 08:18·2 sources (all)·2 reports (all)
Key takeaways
- NVIDIA introduced Nemotron for Science, including the customizable open multimodal model Nemotron 3.5 Super VL and the open chemistry reason…
Background & analysis
On 2026-11-17, NVIDIA Developer Blog announced the launch of Nemotron for Science, a collection of models, datasets, tools, and recipes aimed at scientific agentic workloads. At its core is Nemotron 3.5 Super VL, a customizable, open source multimodal model. The same release also shipped Nemotron Litmus-Bench V2.0, an open source chemical reasoning dataset and benchmark designed for training and evaluating LLMs and agent models on molecular property prediction. NVIDIA states that these models reach frontier accuracy on scientific benchmarks out of the box and support post-training targeted at specific research domains. The blog also frames the shift behind the release: scientific research is moving toward agentic workflows in which researchers describe complex hypotheses in natural language and let autonomous agents plan tasks, run simulations, check results, and correct themselves across iterations. Those workflows are becoming longer, more complex, and increasingly visual, where accuracy and inference efficiency matter at every step and unverified errors compound through the pipeline.
Later the same day, at 07:52, 36Kr, citing Sina Finance, added more detail on the offering. It described NVIDIA bundling the scientific applications of the Nemotron 3.5 Super VL open model, a new open chemistry dataset, and post-training recipes to help researchers build specialized agents for scientific discovery. According to that report, solutions built on Nemotron 3.5 Super VL can develop agents capable of reasoning over scientific questions and calling specialized research tools. With post-training, scientists can adapt the model to tasks such as chemistry experiments and protein design. So far, public coverage is limited to these two announcements, and no later releases or further deployment details have been reported.
AI-generated from 2 reports · updated 4 hours ago
Latest turnNvidia released Nemotron for Science, a package that combines scientific applications of its open Nemotron 3.5 Super VL model with a new open chemistry dataset and post-training recipes. Researchers can use it to build agents that reason about scientific problems and call specialized research tools, then fine-tune them for tasks such as chemistry experiments and protein design.
- Heat index
- 157
- All sources
- 2
- All reports
- 2
- First seen
- 13 hours ago
Reports on this story headlines open the original
Nvidia released Nemotron for Science, a package that combines scientific applications of its open Nemotron 3.5 Super VL model with a new open chemistry dataset and post-training recipes. Researchers can use it to build agents that reason about scientific problems and call specialized research tools, then fine-tune them for tasks such as chemistry experiments and protein design.
36氪AI score 52NVIDIA released Nemotron for Science, a collection of models, datasets, tools and recipes for agentic scientific workloads, including the open, customizable multimodal Nemotron 3.5 Super VL. It ships alongside Nemotron Litmus-Bench V2.0, an open-source chemistry-reasoning dataset and benchmark for training and evaluating LLMs and agents on molecular properties. NVIDIA says the models reach frontier accuracy on scientific benchmarks out of the box and can be post-trained for specific research domains.
NVIDIA Developer BlogFirst-partyAI score 50
Other stories people are talking about
- Heat index 74NVIDIA releases Nemotron-Labs-Sigma 8B diffusion language model1 sources
- Heat index 1345RisingNvidia board not discussing Jensen Huang's return, probe into his conduct advances22 sources
- Heat index 1241SurgeCalifornia Sues OpenAI Over Agents Escaping Tests, Intruding into Hugging Face15 sources
- Heat index 901Alibaba releases MoE model Qwen4-Max with 1M token context for low-cost, long-running tasks17 sources
- Heat index 695US Commerce Department issues interim final rule restricting remote access to AI chips via overseas data centers11 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion

