DEVELOPER REFERENCE · VERIFIED MODEL EVIDENCE

Compare 49 source-verified model records, repository licenses, published weight formats and dated evidence. No overall model ranking.

Evidence statusSource-first verified
Latest record check2026-11-16
Latest record full verification2026-11-16

DIRECT ANSWER · SOURCE-OBSERVED REGISTRY

What NVIDIA model evidence is available?

Open Model Weights currently tracks 49 field-verified NVIDIA model records. In this indexed sample, the recorded modalities are 17 Text; 4 Text → text; 4 Text, image, video → text; 3 Any To Any; 3 Audio → text; 2 Audio To Audio; 2 Embeddings / features; 2 Image + text → text; 1 Action-conditioned driving sensor simulation; 1 Audio and text instructions → text and word timestamps; 1 Document images → OCR text, layout and bounding boxes; 1 Multi-view camera and LiDAR world simulation; 1 Not declared; 1 Query and text/image candidate → relevance score; 1 Simulator-conditioned video generation; 1 Text and images → text; 1 Text assistant and tool use; 1 Text or document images → dense retrieval embedding; 1 Text Ranking; 1 Two-arm robot action to visual observation. The published repository license labels include 20 other; 13 NVIDIA OpenMDW License Agreement v1.1; 5 apache-2.0; 4 mit; 4 OpenMDW 1.1; 2 Unknown / not declared; 1 cc-by-4.0. The listed weight formats include Safetensors (44), PyTorch (5), GGUF (4).

Coverage describes the model repositories indexed here, not every model released by the publisher. A checked source, an unknown field and an unchanged source revision are different states. No missing value is treated as a negative answer.

49verified model records
20recorded modalities
7distinct license labels
49model repository links

DEVELOPER HQ / PRIMARY ORGANIZATION LOCATION

Primary organization location

Santa Clara, California, United States

NVIDIA

This location describes the developer or primary organization. It is not the nationality of a model, dataset, contributor, training run or infrastructure.

SOURCE-BACKED PUBLISHER PROFILE

NVIDIA: models, research and access

NVIDIA publishes the Nemotron family, speech and visual-reasoning models, and optimized or quantized versions of partner models. Its official repositories distinguish base checkpoints, NVFP4/FP8 variants, specialized tasks and model-specific terms.

NVIDIA's verified repository records are not all models trained from scratch by NVIDIA. Several are optimized versions of other developers' checkpoints. Performance, inherited model rights, inference containers and weight licenses have different provenance and must not be conflated.

01

Nemotron models and quantized variants

The tracked Nemotron 3 entries cover language generation, multimodal processing and several quantizations. A BF16 checkpoint and an NVFP4 variant are different artifacts: weight representation can affect memory but does not by itself prove throughput or response quality on a given GPU.

02

Beyond LLMs: Cosmos and speech

The repository index also includes Cosmos visual-reasoning records, Parakeet automatic speech recognition and multimodal Nemotron models. Use task-specific evaluation criteria and verify additional processors, encoders and audio pipeline components before drawing any suitability conclusion.

03

Partner-derived optimized checkpoints

NVIDIA distributes quantized versions of Gemma and Qwen models in its own repository namespace. A quantized artifact can preserve upstream model attribution while having different file formats, precision and installation constraints. Check both upstream licensing requirements and the exact NVIDIA release card.

04

License terms vary by product layer

The NVIDIA repositories include proprietary model-license labels, Apache-2.0, MIT and Creative Commons signals. NeMo Retriever's own documentation distinguishes its Apache-licensed software from separate NVIDIA NIM container terms, and notes that some model files have additional inherited Llama restrictions. Never infer a container license from the weight label.

Publisher background is attributed to official sources. Model counts and license classifications come from OMW repository evidence; no model benchmark or processing location is inferred.

MODEL-BY-MODEL REFERENCE

Which NVIDIA model record fits your workflow?

Browse by published modality and licensing evidence. These are navigation cues, not quality, speed or suitability rankings.

Image + text → textother

Cosmos-Reason2-2B

Visual reasoning

Inspect the Cosmos record when working with visual understanding rather than generic text chat.

Commercial use: unknownLicense evidence →

RELATED RESEARCH & RELEASES

Source-backed publisher coverage

News about additional products is not automatically a field-verified registry addition.

FIELD-BY-FIELD COMPARISON · PRIMARY LINKS

Model evidence at a glance.

Compare the exact entries in our registry. Parameter values are taken from the source-checked record; a license label is not a legal opinion, and recognized weight files are not an inference-speed benchmark. Open each record for the underlying citations and observed history.

Model and parameter countModalityLicense evidenceFormats and filesSource and history
Qwen3.6-35B-A3B-NVFP418.6839B published parametersTextapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors4 filesRepository ↗Model evidence →History →
NVIDIA-Nemotron-3-Nano-4B-BF163.97356B published parametersTextotherunknown commercial-use classificationLicense file ↗Safetensors1 filesRepository ↗Model evidence →History →
bigvgan_v2_22khz_80band_256xPending verification published parametersAudio To Audiomitallowed commercial-use classificationLicense file ↗PyTorch4 filesRepository ↗Model evidence →History →
Gemma-4-26B-A4B-NVFP414.3869B published parametersTextapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →
bigvgan_v2_44khz_128band_512xPending verification published parametersAudio To Audiomitallowed commercial-use classificationLicense file ↗PyTorch4 filesRepository ↗Model evidence →History →
NVIDIA Nemotron 3 Nano 30B-A3B31B total / ~3B active published parametersTextotherunknown commercial-use classificationInspect license evidence →Safetensors14 filesRepository ↗Model evidence →History →
Gemma-4-31B-IT-NVFP420.8686B published parametersTextotherunknown commercial-use classificationInspect license evidence →Safetensors5 filesRepository ↗Model evidence →History →
Qwen2.5-VL-7B-Instruct-NVFP45.02952B published parametersTextotherunknown commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →
NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16560.525B published parametersTextotherunknown commercial-use classificationInspect license evidence →Safetensors226 filesRepository ↗Model evidence →History →
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF1631.5779B published parametersTextotherunknown commercial-use classificationLicense file ↗Safetensors15 filesRepository ↗Model evidence →History →
NVIDIA-Nemotron-3-Super-120B-A12B-BF16123.611B published parametersTextotherunknown commercial-use classificationInspect license evidence →Safetensors51 filesRepository ↗Model evidence →History →
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP417.8202B published parametersTextotherunknown commercial-use classificationLicense file ↗Safetensors53 filesRepository ↗Model evidence →History →
NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP418.2378B published parametersTextotherunknown commercial-use classificationInspect license evidence →Safetensors6 filesRepository ↗Model evidence →History →
Cosmos-Reason2-2B2.4387B published parametersImage + text → textotherunknown commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
parakeet-tdt-0.6b-v30.627057B published parametersAudio → textcc-by-4.0unknown commercial-use classificationInspect license evidence →GGUF, Safetensors2 filesRepository ↗Model evidence →History →
nemotron-3.5-asr-streaming-0.6b0.637997B published parametersAudio → textotherunknown commercial-use classificationInspect license evidence →GGUF, Safetensors2 filesRepository ↗Model evidence →History →
NVIDIA-Nemotron-3-Super-120B-A12B-NVFP467.2286B published parametersTextotherunknown commercial-use classificationInspect license evidence →Safetensors18 filesRepository ↗Model evidence →History →
llama-nemotron-rerank-1b-v21.23582B published parametersText Rankingotherunknown commercial-use classificationLicense file ↗PyTorch, Safetensors2 filesRepository ↗Model evidence →History →
Qwen3.8-27B-NVFP418.1646B published parametersTextapache-2.0allowed commercial-use classificationLicense file ↗Safetensors4 filesRepository ↗Model evidence →History →
dragon-multiturn-query-encoderPending verification published parametersEmbeddings / featuresUnknown / not declaredunknown commercial-use classificationInspect license evidence →PyTorch1 filesRepository ↗Model evidence →History →
dragon-multiturn-context-encoderPending verification published parametersEmbeddings / featuresUnknown / not declaredunknown commercial-use classificationInspect license evidence →PyTorch1 filesRepository ↗Model evidence →History →
Qwen3.6-27B-NVFP418.1646B published parametersTextapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors4 filesRepository ↗Model evidence →History →
NVIDIA-Nemotron-Nano-9B-v28.88823B published parametersTextotherunknown commercial-use classificationInspect license evidence →Safetensors5 filesRepository ↗Model evidence →History →
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF1633.0156B published parametersAny To Anyotherunknown commercial-use classificationInspect license evidence →Safetensors18 filesRepository ↗Model evidence →History →
Qwen3.8-Flash-Next-NVFP4119.602B published parametersImage + text → textotherunknown commercial-use classificationInspect license evidence →Safetensors12 filesRepository ↗Model evidence →History →
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-FP833.0137B published parametersAny To Anyotherunknown commercial-use classificationInspect license evidence →Safetensors5 filesRepository ↗Model evidence →History →
nemotron-speech-streaming-en-0.6b0.618085B published parametersAudio → textotherunknown commercial-use classificationInspect license evidence →GGUF, Safetensors2 filesRepository ↗Model evidence →History →
Nemotron 3.5 Super VL NVFP4120B total / 12B active (publisher-declared) published parametersText, image, video → textNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationInspect license evidence →Safetensors89 filesRepository ↗Model evidence →History →
Nemotron 3.5 Super VL BF16120B total / 12B active (publisher-declared) published parametersText, image, video → textNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationInspect license evidence →Safetensors65 filesRepository ↗Model evidence →History →
NVIDIA Nemotron 3.5 Transcribe 30B-A4B34B total / 4.3B active (publisher-declared) published parametersAudio and text instructions → text and word timestampsNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationLicense file ↗Safetensors1 filesRepository ↗Model evidence →History →
Cosmos3-Edge3.859B published parametersNot declaredotherunknown commercial-use classificationInspect license evidence →Safetensors6 filesRepository ↗Model evidence →History →
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP418.3263B published parametersAny To Anyotherunknown commercial-use classificationInspect license evidence →Safetensors4 filesRepository ↗Model evidence →History →
Nemotron Labs Dory 1.0 GGUF Q5_K_M4B (publisher-declared) published parametersText → textNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationInspect license evidence →GGUF1 filesRepository ↗Model evidence →History →
Qwen3.5-122B-A10B-NVFP464.5808B published parametersTextapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors10 filesRepository ↗Model evidence →History →
GLM-5.2-NVFP4380.989B published parametersTextmitallowed commercial-use classificationInspect license evidence →Safetensors48 filesRepository ↗Model evidence →History →
Nemotron 3.5 Bolt E4B BF167.99616B published parametersText assistant and tool useNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →
Nemotron 3.5 Parse 631M0.6308B total (publisher-declared) published parametersDocument images → OCR text, layout and bounding boxesNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationLicense file ↗Safetensors2 filesRepository ↗Model evidence →History →
Nemotron 3.5 Bolt IQ2 24GB74B total / 9B active (publisher-declared) published parametersText, image, video → textNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationInspect license evidence →Safetensors91 filesRepository ↗Model evidence →History →
Nemotron Labs Dory 1.0 BF164B (publisher-declared) published parametersText → textNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
DeepSeek V4.1 Flash 1017 NVFP4Mixed-precision NVIDIA quantized derivative; do not infer a new base-model parameter total from packed weight tensors published parametersText and images → textmitallowed commercial-use classificationLicense file ↗Safetensors49 filesRepository ↗Model evidence →History →
Nemotron 3.5 Embed 8B BF168B total (publisher-declared) published parametersText or document images → dense retrieval embeddingNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationInspect license evidence →Safetensors5 filesRepository ↗Model evidence →History →
Nemotron 3.5 Bolt BF1674B total / 9B active (publisher-declared) published parametersText, image, video → textNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationLicense file ↗Safetensors64 filesRepository ↗Model evidence →History →
Nemotron Labs Dory 1.0 NVFP44B (publisher-declared) published parametersText → textNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
Nemotron 3.5 Rerank 8B BF168B total (publisher-declared) published parametersQuery and text/image candidate → relevance scoreNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationLicense file ↗Safetensors6 filesRepository ↗Model evidence →History →
Cosmos3-Nano-Sim-Auto15.157B published parametersAction-conditioned driving sensor simulationOpenMDW 1.1allowed-with-conditions commercial-use classificationLicense file ↗Safetensors10 filesRepository ↗Model evidence →History →
Cosmos3-Nano-Sim-Transfer15.156B published parametersSimulator-conditioned video generationOpenMDW 1.1allowed-with-conditions commercial-use classificationLicense file ↗Safetensors9 filesRepository ↗Model evidence →History →
Cosmos3-Nano-Transfer-Auto15.157B published parametersMulti-view camera and LiDAR world simulationOpenMDW 1.1allowed-with-conditions commercial-use classificationLicense file ↗Safetensors11 filesRepository ↗Model evidence →History →
Cosmos3-Nano-Sim-Bimanual15.173B published parametersTwo-arm robot action to visual observationOpenMDW 1.1allowed-with-conditions commercial-use classificationLicense file ↗Safetensors9 filesRepository ↗Model evidence →History →
Nemotron Labs Dory 1.0 MXFP84B (publisher-declared) published parametersText → textNVIDIA OpenMDW License Agreement v1.1allowed-with-conditions commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →

LICENSE & COVERAGE · NOT LEGAL ADVICE

How do the recorded licenses differ?

Different models from the same developer can use different license terms. The distribution below describes the labels and field-level classifications observed in the listed repositories; it does not automatically establish permissions for a specific deployment.

License labels

  • other20 of 49
  • NVIDIA OpenMDW License Agreement v1.113 of 49
  • apache-2.05 of 49
  • mit4 of 49
  • OpenMDW 1.14 of 49
  • Unknown / not declared2 of 49
  • cc-by-4.01 of 49

Commercial-use classifications

  • unknown23 of 49
  • allowed-with-conditions17 of 49
  • allowed9 of 49

Model modalities

  • Text17 of 49
  • Text → text4 of 49
  • Text, image, video → text4 of 49
  • Any To Any3 of 49
  • Audio → text3 of 49
  • Audio To Audio2 of 49
  • Embeddings / features2 of 49
  • Image + text → text2 of 49
  • Action-conditioned driving sensor simulation1 of 49
  • Audio and text instructions → text and word timestamps1 of 49
  • Document images → OCR text, layout and bounding boxes1 of 49
  • Multi-view camera and LiDAR world simulation1 of 49
  • Not declared1 of 49
  • Query and text/image candidate → relevance score1 of 49
  • Simulator-conditioned video generation1 of 49
  • Text and images → text1 of 49
  • Text assistant and tool use1 of 49
  • Text or document images → dense retrieval embedding1 of 49
  • Text Ranking1 of 49
  • Two-arm robot action to visual observation1 of 49

A classification of unknown is not a prohibition. Even a recognized permissive license can carry conditions. Review the license for the exact model and revision. Open the license explorer →

EVIDENCE METHOD · TRANSPARENT LIMITS

How this developer profile is assembled.

We aggregate the current source-checked model records, rather than publish an unverified corporate biography. Repository metadata, model cards, license files and recognized weight artifacts supply the facts where available. The record-check date does not assert that a model or its license changed on that date.

Source-derived runtime signals are not hardware benchmarks. Weight-only memory estimates omit runtime overhead, activations and other deployment requirements. Model-specific uncertainties remain visible in the individual evidence records.

QUESTIONS · DIRECT ANSWERS

Frequently asked questions about NVIDIA models.

Which NVIDIA models are covered?

Open Model Weights currently lists 49 field-verified records for NVIDIA. Their recorded model modalities are: 17 Text; 4 Text → text; 4 Text, image, video → text; 3 Any To Any; 3 Audio → text; 2 Audio To Audio; 2 Embeddings / features; 2 Image + text → text; 1 Action-conditioned driving sensor simulation; 1 Audio and text instructions → text and word timestamps; 1 Document images → OCR text, layout and bounding boxes; 1 Multi-view camera and LiDAR world simulation; 1 Not declared; 1 Query and text/image candidate → relevance score; 1 Simulator-conditioned video generation; 1 Text and images → text; 1 Text assistant and tool use; 1 Text or document images → dense retrieval embedding; 1 Text Ranking; 1 Two-arm robot action to visual observation. This is a registry sample, not an exhaustive publisher catalog.

Which licenses are recorded for NVIDIA models?

In the 49 indexed records, repository license labels are: 20 other; 13 NVIDIA OpenMDW License Agreement v1.1; 5 apache-2.0; 4 mit; 4 OpenMDW 1.1; 2 Unknown / not declared; 1 cc-by-4.0. License labels do not replace the model-specific terms or establish rights for every use case.

Where can I verify NVIDIA model information?

Each indexed model links to its official source repository and a field-level Open Model Weights profile. The most recent record checks shown here reach 2026-11-16; checking a repository does not imply that its contents changed.

Are all NVIDIA models on this page originally trained by NVIDIA?

No. Some are NVIDIA-branded Nemotron or Cosmos checkpoints; others are optimized versions of models originally published by partners such as Google and Qwen. Inspect the base-model and publisher chain.

Does NVFP4 guarantee faster inference than BF16?

No. NVFP4 denotes an artifact representation, not an independently measured speedup. Serving-stack compatibility, kernels, accelerator type, batch size and model architecture determine results.

Do NVIDIA NIM container terms match every open-weight license?

No. NVIDIA documents separate terms for NIM images and model components. Users must inspect all applicable weight, software and redistribution licenses for their selected deployment.

COMPLETE INDEXED SELECTION

All 49 indexed model profiles.

Qwen3.6-35B-A3B-NVFP4
Text
262,144 tokens
apache-2.0
Field-verified
NVIDIA-Nemotron-3-Nano-4B-BF16
Text
262,144 tokens
other
Field-verified
bigvgan_v2_22khz_80band_256x
Audio To Audio
Not declared in checked config/model card
mit
Field-verified
Gemma-4-26B-A4B-NVFP4
Text
262,144 tokens
apache-2.0
Field-verified
bigvgan_v2_44khz_128band_512x
Audio To Audio
Not declared in checked config/model card
mit
Field-verified
NVIDIA Nemotron 3 Nano 30B-A3B
Text
262,144 tokens
other
Field-verified
Gemma-4-31B-IT-NVFP4
Text
262,144 tokens
other
Field-verified
Qwen2.5-VL-7B-Instruct-NVFP4
Text
128,000 tokens
other
Field-verified
NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16
Text
262,144 tokens
other
Field-verified
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16
Text
262,144 tokens
other
Field-verified
NVIDIA-Nemotron-3-Super-120B-A12B-BF16
Text
262,144 tokens
other
Field-verified
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4
Text
1,048,576 tokens
other
Field-verified
NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP4
Text
262,144 tokens
other
Field-verified
Cosmos-Reason2-2B
Image + text → text
Not declared in checked config/model card
other
Field-verified
parakeet-tdt-0.6b-v3
Audio → text
5,000 tokens
cc-by-4.0
Field-verified
nemotron-3.5-asr-streaming-0.6b
Audio → text
5,000 tokens
other
Field-verified
NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4
Text
262,144 tokens
other
Field-verified
llama-nemotron-rerank-1b-v2
Text Ranking
131,072 tokens
other
Field-verified
Qwen3.8-27B-NVFP4
Text
262,144 tokens
apache-2.0
Field-verified
dragon-multiturn-query-encoder
Embeddings / features
512 tokens
Unknown / not declared
Field-verified
dragon-multiturn-context-encoder
Embeddings / features
512 tokens
Unknown / not declared
Field-verified
Qwen3.6-27B-NVFP4
Text
262,144 tokens
apache-2.0
Field-verified
NVIDIA-Nemotron-Nano-9B-v2
Text
131,072 tokens
other
Field-verified
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16
Any To Any
131,072 tokens
other
Field-verified
Qwen3.8-Flash-Next-NVFP4
Image + text → text
262,144 tokens
other
Field-verified
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-FP8
Any To Any
131,072 tokens
other
Field-verified
nemotron-speech-streaming-en-0.6b
Audio → text
5,000 tokens
other
Field-verified
Nemotron 3.5 Super VL NVFP4
Text, image, video → text
1,048,576 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified
Nemotron 3.5 Super VL BF16
Text, image, video → text
1,048,576 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified
NVIDIA Nemotron 3.5 Transcribe 30B-A4B
Audio and text instructions → text and word timestamps
262,144 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified
Cosmos3-Edge
Not declared
131,072 tokens
other
Field-verified
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4
Any To Any
131,072 tokens
other
Field-verified
Nemotron Labs Dory 1.0 GGUF Q5_K_M
Text → text
Not declared in checked config/model card
NVIDIA OpenMDW License Agreement v1.1
Field-verified
Qwen3.5-122B-A10B-NVFP4
Text
262,144 tokens
apache-2.0
Field-verified
GLM-5.2-NVFP4
Text
1,048,576 tokens
mit
Field-verified
Nemotron 3.5 Bolt E4B BF16
Text assistant and tool use
65,536 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified
Nemotron 3.5 Parse 631M
Document images → OCR text, layout and bounding boxes
16,384 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified
Nemotron 3.5 Bolt IQ2 24GB
Text, image, video → text
262,144 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified
Nemotron Labs Dory 1.0 BF16
Text → text
262,144 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified
DeepSeek V4.1 Flash 1017 NVFP4
Text and images → text
1,048,576 tokens
mit
Field-verified
Nemotron 3.5 Embed 8B BF16
Text or document images → dense retrieval embedding
262,144 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified
Nemotron 3.5 Bolt BF16
Text, image, video → text
262,144 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified
Nemotron Labs Dory 1.0 NVFP4
Text → text
262,144 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified
Nemotron 3.5 Rerank 8B BF16
Query and text/image candidate → relevance score
262,144 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified
Cosmos3-Nano-Sim-Auto
Action-conditioned driving sensor simulation
Not verified from source configuration
OpenMDW 1.1
Field-verified
Cosmos3-Nano-Sim-Transfer
Simulator-conditioned video generation
Not verified from source configuration
OpenMDW 1.1
Field-verified
Cosmos3-Nano-Transfer-Auto
Multi-view camera and LiDAR world simulation
Not verified from source configuration
OpenMDW 1.1
Field-verified
Cosmos3-Nano-Sim-Bimanual
Two-arm robot action to visual observation
Not verified from source configuration
OpenMDW 1.1
Field-verified
Nemotron Labs Dory 1.0 MXFP8
Text → text
262,144 tokens
NVIDIA OpenMDW License Agreement v1.1
Field-verified