# NVIDIA Introduces Nemotron for Science to Accelerate Discoveries With Agents

[Canonical reading](https://escapemeridian.com/readings/nvidia-introduces-nemotron-for-science-to-accelerate-c70c9444)

NVIDIA says Nemotron for Science brings together its Nemotron 3.5 Super VL open model, the LitmusBench v0.2 chemistry dataset and post-training recipes so researchers can build specialized agents for scientific work. Post-training means adapting a model with domain-specific data. The announcement describes agents that can interpret scientific information, generate code and use computational tools. NVIDIA reports examples from partner companies and research institutions, including extracting reaction details from diagrams, designing laboratory protocols and repairing GROMACS molecular dynamics setups. These results are tied to particular tests or workflows. NVIDIA’s announcement says Terray Therapeutics’ model surpassed previously published results on two benchmarks, and reports that KTH’s agent fixed 81% of test cases in an overnight run, with every fix checked by GROMACS. NVIDIA describes LitmusBench v0.2 as a dataset for training and evaluating chemistry reasoning, and says the model and dataset are available on Hugging Face. Escape Meridian interpretation: these examples illustrate the scientific tasks NVIDIA and its collaborators are pursuing, but do not establish performance across other research settings. The results and claims are reported in NVIDIA’s announcement; the supplied text provides no independent evaluation of them.

Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.

## Original source

NVIDIA · Article

Original source date: Nov 16, 2026 (2026-11-16T16:00:00.000Z).

Date basis: Date supplied by the publisher’s feed.

Added to Escape Meridian: 2026-11-16T16:32:07.052Z. First observed: 2026-11-16T16:06:46.033Z. Retrieved: 2026-11-16T16:19:25.660Z.

- [NVIDIA Introduces Nemotron for Science to Accelerate Discoveries With Agents](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Limits

The evidence is NVIDIA's announcement; the reported partner results and claims are not independently evaluated in the supplied text. The performance examples concern specific benchmarks or workflows, and the announcement does not establish how results generalize to other scientific settings.


## Open question

How do the reported results hold up on scientific tasks beyond the specific benchmarks and workflows described in the announcement?


## Source-stated

NVIDIA presents Nemotron for Science as a collection of an open model, a chemistry dataset and post-training recipes intended to help researchers build specialized science agents.

> NVIDIA Nemotron for Science, bringing together scientific applications of its NVIDIA Nemotron™ 3.5 Super VL open model, a new open chemistry dataset and post-training recipes

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Source-stated

NVIDIA says the agents can be adapted to specialized scientific tasks and can use research tools in workflows.

> scientists can adapt the model to specialized tasks such as conducting experiments for chemistry, protein design and more.

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Source-stated

NVIDIA reports that Terray Therapeutics' post-trained model surpassed previously published results on two benchmarks for extracting reaction information from scientific diagrams.

> surpassing previously published results on two benchmarks. On the RxnScribe test, it recovered about 76% of annotated reactions

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Source-stated

NVIDIA reports that KTH's agent fixed 81% of test cases in one overnight run; GROMACS independently checked every fix.

> In one overnight run on a single NVIDIA DGX Spark™ system, the agent fixed 81% of test cases with reasoning enabled, and every fix was checked independently by GROMACS itself.

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Source-stated

NVIDIA says Nemotron 3.5 Super VL and LitmusBench v0.2 are available on Hugging Face.

> NVIDIA Nemotron 3.5 Super VL and the LitmusBench v0.2 chemistry dataset are available now on Hugging Face.

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Escape Meridian interpretation

The announcement's reported results concern particular benchmarks and workflows, so they do not by themselves establish performance across other research settings.

> Terray Therapeutics post-trained Nemotron to extract reactants, products and reaction conditions from scientific diagrams, surpassing previously published results on two benchmarks.

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Source and contributor context

NVIDIA · Publisher

The publisher is identified here. No individual author has been inferred.


## Source media

Media declared by the original publisher; availability checked separately from the editorial reading.

- [NVIDIA Nemotron for Science · NVIDIA](https://iprsoftwaremedia.com/219/files/202611/bbee83cc9133eb5438c52eab3fd409e2/6af8bac03d6332a556233545_nvidia-nemotron-for-science/nvidia-nemotron-for-science_3ae01a42-fd11-4792-b9a1-a25649fd94c1-prv.jpg?v=3ae01a42-fd11-4792-b9a1-a25649fd94c1)

## Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.


## Benchmark

A defined set of tests used to compare systems. Results depend on the tasks and measures selected.

- [Definition reference](https://arxiv.org/abs/2211.09110)

## NVIDIA

NVIDIA develops processors, networking and software used for AI, graphics and scientific computing.

- [Explore this context](/entities/nvidia)
- [Reference](https://blogs.nvidia.com/)

## Hugging Face

Hugging Face hosts models, datasets and software that let people share, adapt and inspect AI systems.

- [Explore this context](/entities/hugging-face)
- [Reference](https://huggingface.co/blog)

## Editorial provenance

Version 1. Publication ID: 97cc1410-8173-43ee-8378-30a7f3e3ba45.

Prepared by gpt-6-luna; grounding review by gpt-6-luna.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-16T16:32:07.052Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: material\_development.

Observation: 6ffd3d0a8fa1cc562626a6269a5334afa670399aa8673273848d02b28042f5a5. Source text hash: 94fb5c8bbd7436e5b61492a0c76eac0dc1ab503a2abf340ce142c5075d441e98.


## Revision and source-check history

### Version 1 — 2026-11-16T16:32:07.052Z

NVIDIA says Nemotron for Science brings together its Nemotron 3.5 Super VL open model, the LitmusBench v0.2 chemistry dataset and post-training recipes so researchers can build specialized agents for scientific work. Post-training means adapting a model with domain-specific data. The announcement describes agents that can interpret scientific information, generate code and use computational tools. NVIDIA reports examples from partner companies and research institutions, including extracting reaction details from diagrams, designing laboratory protocols and repairing GROMACS molecular dynamics setups. These results are tied to particular tests or workflows. NVIDIA’s announcement says Terray Therapeutics’ model surpassed previously published results on two benchmarks, and reports that KTH’s agent fixed 81% of test cases in an overnight run, with every fix checked by GROMACS. NVIDIA describes LitmusBench v0.2 as a dataset for training and evaluating chemistry reasoning, and says the model and dataset are available on Hugging Face. Escape Meridian interpretation: these examples illustrate the scientific tasks NVIDIA and its collaborators are pursuing, but do not establish performance across other research settings. The results and claims are reported in NVIDIA’s announcement; the supplied text provides no independent evaluation of them.

## Original source

NVIDIA · Article

Original source date: Nov 16, 2026 (2026-11-16T16:00:00.000Z).

Date basis: Date supplied by the publisher’s feed.

Added to Escape Meridian: 2026-11-16T16:32:07.052Z. First observed: 2026-11-16T16:06:46.033Z. Retrieved: 2026-11-16T16:19:25.660Z.

- [NVIDIA Introduces Nemotron for Science to Accelerate Discoveries With Agents](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Limits

The evidence is NVIDIA's announcement; the reported partner results and claims are not independently evaluated in the supplied text. The performance examples concern specific benchmarks or workflows, and the announcement does not establish how results generalize to other scientific settings.


## Open question

How do the reported results hold up on scientific tasks beyond the specific benchmarks and workflows described in the announcement?


## Source-stated

NVIDIA presents Nemotron for Science as a collection of an open model, a chemistry dataset and post-training recipes intended to help researchers build specialized science agents.

> NVIDIA Nemotron for Science, bringing together scientific applications of its NVIDIA Nemotron™ 3.5 Super VL open model, a new open chemistry dataset and post-training recipes

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Source-stated

NVIDIA says the agents can be adapted to specialized scientific tasks and can use research tools in workflows.

> scientists can adapt the model to specialized tasks such as conducting experiments for chemistry, protein design and more.

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Source-stated

NVIDIA reports that Terray Therapeutics' post-trained model surpassed previously published results on two benchmarks for extracting reaction information from scientific diagrams.

> surpassing previously published results on two benchmarks. On the RxnScribe test, it recovered about 76% of annotated reactions

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Source-stated

NVIDIA reports that KTH's agent fixed 81% of test cases in one overnight run; GROMACS independently checked every fix.

> In one overnight run on a single NVIDIA DGX Spark™ system, the agent fixed 81% of test cases with reasoning enabled, and every fix was checked independently by GROMACS itself.

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Source-stated

NVIDIA says Nemotron 3.5 Super VL and LitmusBench v0.2 are available on Hugging Face.

> NVIDIA Nemotron 3.5 Super VL and the LitmusBench v0.2 chemistry dataset are available now on Hugging Face.

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Escape Meridian interpretation

The announcement's reported results concern particular benchmarks and workflows, so they do not by themselves establish performance across other research settings.

> Terray Therapeutics post-trained Nemotron to extract reactants, products and reaction conditions from scientific diagrams, surpassing previously published results on two benchmarks.

- [Supporting source](https://nvidianews.nvidia.com/news/nvidia-introduces-nemotron-for-science-to-accelerate-discoveries-with-agents)

## Source and contributor context

NVIDIA · Publisher

The publisher is identified here. No individual author has been inferred.


## Source media

Media declared by the original publisher; availability checked separately from the editorial reading.

- [NVIDIA Nemotron for Science · NVIDIA](https://iprsoftwaremedia.com/219/files/202611/bbee83cc9133eb5438c52eab3fd409e2/6af8bac03d6332a556233545_nvidia-nemotron-for-science/nvidia-nemotron-for-science_3ae01a42-fd11-4792-b9a1-a25649fd94c1-prv.jpg?v=3ae01a42-fd11-4792-b9a1-a25649fd94c1)

## Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.


## Benchmark

A defined set of tests used to compare systems. Results depend on the tasks and measures selected.

- [Definition reference](https://arxiv.org/abs/2211.09110)

## NVIDIA

NVIDIA develops processors, networking and software used for AI, graphics and scientific computing.

- [Explore this context](/entities/nvidia)
- [Reference](https://blogs.nvidia.com/)

## Hugging Face

Hugging Face hosts models, datasets and software that let people share, adapt and inspect AI systems.

- [Explore this context](/entities/hugging-face)
- [Reference](https://huggingface.co/blog)

## Editorial provenance

Version 1. Publication ID: 97cc1410-8173-43ee-8378-30a7f3e3ba45.

Prepared by gpt-6-luna; grounding review by gpt-6-luna.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-16T16:32:07.052Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: material\_development.

Observation: 6ffd3d0a8fa1cc562626a6269a5334afa670399aa8673273848d02b28042f5a5. Source text hash: 94fb5c8bbd7436e5b61492a0c76eac0dc1ab503a2abf340ce142c5075d441e98.


### passed — 2026-11-16T16:32:07.052Z — anthropic/claude-opus-5.5

Publication: 97cc1410-8173-43ee-8378-30a7f3e3ba45

attribution — pass: The reading attributes the product description, partner results and availability to NVIDIA's announcement. It notes that the supplied text gives no independent evaluation. The Escape Meridian interpretation is labeled as such.

> NVIDIA today introduced NVIDIA Nemotron for Science

summary — pass: It leads with the actor and the development, then explains post-training and what the agents do. The examples and the interpretation are accurate. GROMACS and Hugging Face are named without explanation, but they are not central, so a reader can still follow the reading. Omitting the Lila, Therna and Argonne examples is a selection choice, not a distortion. The question is a real open issue that follows from the benchmark-specific results.

> help researchers build specialized agents for scientific discovery

date — pass: originalPublished is 2026-11-16, taken from the feed entry. This matches the feed date and the announcement's 'today' framing. The text shows no earlier first-publication date.

> NVIDIA today introduced NVIDIA Nemotron for Science

support — pass: The Terray benchmark, KTH 81% overnight run with GROMACS verification, Hugging Face availability and LitmusBench description all match the source. Claim 2 attributes adaptation to 'the agents'. The source says the model is adapted, and its excerpt does not cover tool use, which appears elsewhere in the source. This is a minor imprecision, not a material one. The interpretation is labeled and conditional, and it does not overreach.

> In one overnight run on a single NVIDIA DGX Spark™ system, the agent fixed 81% of test cases with reasoning enabled, and every fix was checked independently by GROMACS itself.

media — pass: The selected media is the verified direct link to the original NVIDIA news page. It is the most direct way for a reader to access the source.

## Recorded references

Explicit citations in public readings to this reading. Shared topics are not citations. Other record types need an explicit reference adapter.

No public references recorded here yet.

[Reference data (JSON)](/readings/c70c94442ecac4abaefe235fc5b239f4535020f73cb4eaf74e9f4addafe049f5/references.json)

## Corrections and challenges

[Send a correction or an opposing source](https://escapemeridian.com/corrections/)

[JSON record](https://escapemeridian.com/readings/nvidia-introduces-nemotron-for-science-to-accelerate-c70c9444.json)
