Model release

NVIDIA Nemotron Labs Dory1.0-4B

NVIDIA Nemotron Labs Dory1.0-4B is an open-weight model release in the NVIDIA Nemotron family from NVIDIA1, version Dory1.0-4B, GA (10/16/2026), released Oct 16, 2026. Its weights can be downloaded by the public1 2 3 4 5 under the OpenMDW License Agreement, version 1.11 2 6.

Dory1.0-4B is a 4B-parameter looped (latent-recurrent) language model from NVIDIA's Nemotron Labs. It reuses a shared block of layers several times per token, so one checkpoint can run at different compute levels by changing the loop count. The card lists English as the supported language and a context of up to 256K tokens, and NVIDIA publishes BF16, MXFP8, NVFP4, and GGUF (Q5_K_M) versions.1 2 3 4 5

Model hub: Model card (BF16) (external site: huggingface.co)
Model hub: MXFP8 checkpoint (external site: huggingface.co)
Model hub: NVFP4 checkpoint (external site: huggingface.co)
Model hub: GGUF (Q5_K_M) export (external site: huggingface.co)
License: OpenMDW License Agreement, version 1.1 (external site: openmdw.ai)

On AI on Hugging Face: nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16, nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF, nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-NVFP4, nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-MXFP8

Last reviewedOct 18, 2026Entry updated Oct 18, 2026Documented release Oct 16, 2026

Availability and license

Overall availability

Public

Weights are downloadable from Hugging Face under OpenMDW-1.1; the repository metadata for the BF16, MXFP8, NVFP4, and GGUF repositories showed no access gate when checked.1 2 3 4 5

Availability fact review: not separately recorded

Availability is separate from permission: read the license before using or redistributing.

OpenMDW License Agreement, version 1.1 (external site: openmdw.ai)1 2 6
Applies to: Weights · custom license (no SPDX identifier) · fact reviewed Oct 18, 2026
Plain-language guide to the OpenMDW License Agreement, version 1.1

The card's license field and its terms-of-use section name OpenMDW-1.1; the BF16 repository has no separate LICENSE file. OpenMDW-1.1 permits use, modification, and sharing of the model materials without restriction, including under copyright, patent, database, and trade secret rights. Redistributors must include the agreement and keep applicable origin notices, and rights end for anyone who brings a patent or copyright suit over the materials unless responding to one filed first. It places no conditions on outputs and disclaims warranties. SPDX list version 3.29.0 includes OpenMDW-1.0 but not 1.1, so no SPDX ID is recorded.1 2 6 7

Component reuse rights

A readable or downloadable component is not automatically reusable. These indicators concern recorded license evidence, not system certification.

weights: Unknown — no complete fact-level rights review
code: Unknown — no complete fact-level rights review
data: Unknown — no complete fact-level rights review
documentation: Unknown — no complete fact-level rights review

No complete system-rights review is recorded for this release.

Model-disclosure tier

Computed from the checklist below using USASI rubric v0.2. An editorial category, not a certification.

Open-weight

The model parameters for this release can be downloaded by the public. License terms may still restrict use, redistribution, or commercial use.

The weights are under a license that is not on the rubric's OSI-approved list. Read its terms before use.

How tiers are computed

Public materials checklist

Items for a model under USASI rubric v0.2. Unknown means unassessed or insufficient evidence.

Public materials checklist for NVIDIA Nemotron Labs Dory1.0-4B
Item	Status	Notes and evidence
WeightsCan the general public download the model parameters for this release?	Public	BF16, MXFP8, NVFP4, and GGUF (Q5_K_M) checkpoints are published on Hugging Face.1 2 3 4 5
Inference codeIs code for running the model published?	Partial	The card gives serving instructions for vLLM and llama.cpp, but at release support in both came from pull requests not yet in a tagged release of either project.1
Training codeIs the code used to train the model published?	Unknown	The card names Megatron-LM, NeMo RL, and NeMo Gym as the software used, but no Dory-specific training code was found.
Training-data informationDoes the information cover provenance, scope, acquisition, selection, labeling, processing, and where data or alternatives can be obtained? Access alone does not establish completeness.	Partial	The card describes data types, domains, languages, token counts, collection periods, and post-training synthetic-data and filtering methods at a high level, and refers to the Nemotron 3.5 Lightning card for pre-training and RL dataset details.1
Training-data accessCan the training data be obtained? This is independent of information completeness and reuse rights; original unshareable data need not be downloadable.	Unknown	The repository metadata tags NVIDIA's pre-training and post-training dataset collections, but the card does not state which of the data used is downloadable.
Complete training pipelineIs the complete base-training and preprocessing pipeline published, including configuration? Fine-tuning code or an inference SDK alone is insufficient.	Unknown	Not assessed.
Legacy data assessment (v0.1)Historical assessment combining download access and disclosure. Preserved for traceability; excluded from the v0.2 tier calculation. See the new separate assessments above.	Unknown	Not assessed.
Training recipeAre the training configuration and procedure documented in enough detail to follow?	Partial	The card outlines the stages: 6T tokens of pre-training (NVFP4 for roughly the first 4.5T, then BF16), 300B tokens of context extension, SFT, and GRPO reinforcement learning. No configurations are published with it.1
Evaluation materialsPublic = evaluation code or prompts that let others re-run the evaluations are published. Partial = results only.	Partial	The card reports results measured with NeMo Gym and the NeMo Evaluator SDK but does not publish model-specific commands for re-running them.1

What it is useful for

The model card describes it as aimed at general-purpose reasoning on a small footprint, with the loop count adjustable to trade inference compute against accuracy, and says it is ready for commercial or non-commercial use.1

Run and use notes

Documented facts only. No hardware or performance claims are made without a cited source and stated assumptions.

The card documents serving with vLLM (validated at a 256K-token context) and with llama.cpp on Apple Silicon Macs using the Q5_K_M GGUF export, and lists NVIDIA H100, B200, B300, and DGX Spark as test hardware. It recommends two loops as the default and describes a last-loop KV-cache mode that keeps cache size at the one-loop level regardless of loop count.1

Organization context

NVIDIA

Provenance and derivatives

Trained by NVIDIA from pre-training onward; the card and repository metadata name no base model. The MXFP8, NVFP4, and GGUF repositories are quantized versions of the BF16 checkpoint. The card says post-training synthetic data was generated with teacher models, including third-party open models such as GPT-OSS-120B.1 2

Other releases in the NVIDIA Nemotron family

NVIDIA Nemotron 3 Super 120B-A12BModel-disclosure tier (USASI rubric v0.2): Open-weight
NVIDIA Nemotron 3 Ultra 550B-A55BModel-disclosure tier (USASI rubric v0.2): Open-weight
NVIDIA Nemotron 3.5 Lightning 30B-A3BModel-disclosure tier (USASI rubric v0.2): Open-weight
NVIDIA Nemotron 3.5 Super VL 120B-A12BModel-disclosure tier (USASI rubric v0.2): Open-weight

NVIDIA Nemotron family overview

What this catalog does not know

Unknown means the sources reviewed for this record do not document it. It is not evidence that something does not exist.

Training code: unknown.
Training-data access: unknown.
Complete training pipeline: unknown.
Legacy data assessment (v0.1): unknown.

Have a primary source? How to report a correction.

U.S. eligibility

Project eligibility rests on documented governing or maintaining entities, not on contributors.

Eligible · basis: U.S. headquarters

The model card states the model was trained by NVIDIA. NVIDIA Corporation's principal executive offices are in Santa Clara, California, per its Form 10-Q for the quarter ended July 26, 2026.1 8

Assessed Oct 18, 2026

Sources

1.
NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16 model card (external site: huggingface.co)
NVIDIA (Hugging Face) · Model card · published Oct 16, 2026 · accessed Oct 18, 2026
2.
Hugging Face model metadata for nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16 (external site: huggingface.co)
Hugging Face · Other · accessed Oct 18, 2026
3.
Hugging Face model metadata for nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-MXFP8 (external site: huggingface.co)
Hugging Face · Other · accessed Oct 18, 2026
4.
Hugging Face model metadata for nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-NVFP4 (external site: huggingface.co)
Hugging Face · Other · accessed Oct 18, 2026
5.
Hugging Face model metadata for nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF (external site: huggingface.co)
Hugging Face · Other · accessed Oct 18, 2026
6.
OpenMDW License Agreement, version 1.1 (OpenMDW-1.1) (external site: openmdw.ai)
OpenMDW · License · accessed Oct 18, 2026
7.
SPDX License List data (licenses.json, version 3.29.0) (external site: raw.githubusercontent.com)
SPDX · Other · published Sep 16, 2026 · accessed Oct 18, 2026
8.
NVIDIA Corporation Form 10-Q for the quarter ended July 26, 2026 (external site: sec.gov)
U.S. Securities and Exchange Commission (EDGAR) · Filing · accessed Oct 18, 2026

Email a correction Report a correction Edit this entry

This listing is not an endorsement, a safety assessment, or a federal approval.

Cite this page

APA style

USASI. (2026, October 18). NVIDIA Nemotron Labs Dory1.0-4B. United States of America Superintelligence. https://unitedstatesofamericasuperintelligence.com/open/nemotron-labs-dory-1-0-4b/

BibTeX

@misc{usasi_nemotron_labs_dory_1_0_4b,
  author = {{USASI}},
  title = {NVIDIA Nemotron Labs Dory1.0-4B},
  year = {2026},
  month = oct,
  howpublished = {\url{https://unitedstatesofamericasuperintelligence.com/open/nemotron-labs-dory-1-0-4b/}},
  note = {United States of America Superintelligence. Last updated 2026-10-18}
}

The date is when this page was last updated. Add the date you read it if your style needs one.

For maintainers: a “Listed in USASI” badge

Links to this page. It says the entry is listed, not endorsed or rated. About the badge

Markdown (for a README)

[![Listed in the USASI catalog](https://unitedstatesofamericasuperintelligence.com/badges/listed.svg)](https://unitedstatesofamericasuperintelligence.com/open/nemotron-labs-dory-1-0-4b/)

HTML

<a href="https://unitedstatesofamericasuperintelligence.com/open/nemotron-labs-dory-1-0-4b/"><img src="https://unitedstatesofamericasuperintelligence.com/badges/listed.svg" alt="Listed in the USASI catalog" height="20"></a>