Model index · NVIDIA · updated 2026-08-16

NVIDIA models and prices

15 models, 9 with a published price , 1 with open weights · released 2025-12-14 to 2026-08-11.

NVIDIA does not sell these models — every price here is a host's price

That is the normal shape for a lab that publishes weights instead of selling inference: the model is theirs, the price is somebody else's. It also means there is no single official price to quote, only a market of hosts.

What NVIDIA makes

CategoryModelsBrowse
Text 10 all Text models →
Video 4 all Video models →
Image 1 all Image models →

Cheapest NVIDIA models per 1M tokens

Standard on-demand tier, cheapest host per model. Media models are priced in their own unit and listed further down rather than mixed into this table.

ModelBlended /1MCheapest hostQualityContext
Nemotron 3.5 Lightning $0.087 DeepInfra not measured 1M
Nemotron 3 Nano 30B A3B $0.087 OpenRouter (default route) * not measured 262k
Nemotron 3 Super $0.164 OpenRouter (default route) * not measured 1M
Nemotron 3 Ultra $0.925 DeepInfra not measured 1M

Open weights from NVIDIA

Parameter counts are computed from the published weight files, and licences are what the authors attached — both from the weights registry rather than from any seller. Licence class decides whether you can self-host commercially without signing something.

ModelParametersLicenceDownloads (30d)Repo
Nemotron 3 Ultra 335.04B openmdw-1.1 0.25M nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4

Who sells NVIDIA models

Offering rows per seller across this vendor's models. A seller appearing many times serves many of the models, not the same one repeatedly.

SellerOfferings
DeepInfra 7
OpenRouter (default route) 5
fal 5
Amazon (Bedrock) 2
Novita 2
CoreWeave 1
BaseTen 1
Venice 1
Together 1
DigitalOcean 1
Nebius 1
Crusoe 1

Jobs these models do

Comparison on each model page is restricted to models doing the same job — a price difference between two different jobs is arithmetic rather than advice.

JobModels
text generation10
image-to-video generation2
text-to-video generation2
text-to-image generation1

One API for every NVIDIA model

VernaOne fronts the sellers on this page behind a single interface, so switching model or host is a config change — with automatic fallback if quality or latency regresses.

Try VernaOne free →

All 15 NVIDIA models

Newest first, priced or not. A model with no price is a fact about what its sellers publish, not a gap we are hiding.

ModelCategoryReleasedPriceKindJob
Nemotron 3.5 Lightning Text 2026-08-11 $0.087 /1M reasoning text generation
Nemotron Asr Multilingual Text 2026-06-10 no published price text generation
Nemotron 3 Ultra Text 2026-06-04 $0.925 /1M reasoning
open weights
text generation
Cosmos 3 Super Image to Video Video 2026-06-02 $0.05 /second of video image-to-video generation
Cosmos 3 Super Image 2026-06-01 $0.02 /request text-to-image generation
Nemotron Diffusion Vlm Text 2026-05-21 no published price text generation
Nemotron 3 Nano Omni (vision) Text 2026-04-28 no published price text generation
Nemotron 3 Nano Omni (video) Text 2026-04-28 no published price text generation
Nemotron 3 Nano Omni (audio) Text 2026-04-28 no published price text generation
Nemotron 3 Nano Omni Text 2026-04-28 no published price text generation
Nemotron 3 Super Text 2026-03-11 $0.164 /1M reasoning text generation
Cosmos Predict 2.5 2B (text to video) Video 2026-02-24 $0.2 /second of video text-to-video generation
Cosmos Predict 2.5 2B (image to video) Video 2026-02-24 $0.2 /second of video image-to-video generation
Cosmos Predict 2.5 2B Distilled Video 2026-02-24 $0.08 /second of video text-to-video generation
Nemotron 3 Nano 30B A3B Text 2025-12-14 $0.087 /1M reasoning text generation

Other companies

Alibaba · Black Forest Labs · LTX · Kling · OpenAI · Google · Stability AI · Minimax · Bria AI · Bytedance · Pixverse · Ideogram · Tencent · Luma AI · All models

Frequently asked

How many models does NVIDIA have?

We track 15 NVIDIA models across 3 categories, 9 of them with a published price. The list is rebuilt from provider APIs, the vendor's own docs and the weights registry rather than maintained by hand, so new releases appear without us writing about them.

Can I buy NVIDIA models directly from NVIDIA?

Not for the models on this page. Every price we hold for NVIDIA comes from a third-party host, which is what you would expect from a lab that publishes weights rather than selling inference.

Are NVIDIA models open weights?

1 of 15 have a public weights repository. Licence mix: 1 source available. Licence class decides whether you can self-host commercially without a separate agreement, so check the model page before planning a deployment.