SAVRN
Search Contact SAVRN

Open-weight model

Cosmos3-Edge

by NVIDIA nvidia/Cosmos3-Edge

NVIDIA Cosmos™ is a world foundation model platform designed to accelerate the development of Physical AI by enabling machines to understand, simulate, and interact with the physical world across robotics, autonomous driving, and smart space environments…

Parameters3.9B
Context131,072
Weights9.1 GB
Licenseother
AccessOpen weights
Monthly Downloads1.5M

Runs On

What it takes to serve Cosmos3-Edge (3.9B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

PrecisionWeightsMemory neededCheapest setupPer hourAlso fits
16-bit 7.7 GB 9.3 GB 1x MI300X (192 GB)
Vultr
$1.85 1x H100 $1.99 · 1x MI325X $2.00
8-bit 3.9 GB 4.6 GB 1x MI300X (192 GB)
Vultr
$1.85 1x H100 $1.99 · 1x MI325X $2.00
4-bit 1.9 GB 2.3 GB 1x MI300X (192 GB)
Vultr
$1.85 1x H100 $1.99 · 1x MI325X $2.00

Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.

SAVRN's Notes on Cosmos3-Edge

Nine point three gigabytes of memory is the whole ask for Cosmos3-Edge at 16-bit, and that number settles the hardware question. NVIDIA built this 3.9 billion parameter world model for physical AI: text, image, video or an action trajectory goes in, and video, image, audio or action commands come out for robots, vehicles and factory floors. At 8-bit the working set drops to 4.6 GB and at 4-bit to 2.3 GB, so the cheapest card on the Index, one MI300X with 192 GB at $1.85 an hour on-demand, carries it with most of the card idle. It belongs beside larger jobs.

The license reads other, with no summary in our record, so pull the actual terms before any commercial deployment. Check the 131,072-token context against your real input lengths, and watch the dates: released July 1, 2026, updated September 16, 2026, so the files are still moving.

Model Card

NVIDIA Cosmos™ is a world foundation model platform designed to accelerate the development of Physical AI by enabling machines to understand, simulate, and interact with the physical world across robotics, autonomous driving, and smart space environments, including industrial and factory-scale applications. Cosmos3 is a collection of Omnimodal world models capable of generating dynamic, high-quality video, image, audio, and action commands from combinations of text, image, video, and action trajectory inputs. It serves as a foundational building block for a broad range of Physical AI applications and research spanning world understanding, world generation, simulation, and embodied policy…

Excerpt from the card by NVIDIA, licensed other.

Configuration

Architecture
Cosmos3EdgeForConditionalGeneration
Context length (tokens)
131,072
Layers
28
Hidden size
2,048
Feed-forward size
9,216
Attention heads
16
Key/value heads
8
Head dimension
128
Vocabulary size
131,072
Model type
cosmos3_edge

Identity and Version

Repository
nvidia/Cosmos3-Edge
Publisher
NVIDIA
Task
Not stated by the source
Modality
Other
Library
cosmos
Parameters
3.9B parameters
Languages
Not stated by the source
Revision
344d602b128d1bbdacb43b08d0a3626f46343e29
First published
2026-07-01
Last updated
2026-09-16

Files and Weights

54 files, 9.2 GB in total. The weights are 4 files totalling 9.1 GB in safetensors.

Weights4 files · 9.1 GB
Configuration22 files · 223.4 KB
Tokenizer4 files · 34.3 MB
Documentation5 files · 65.8 KB
Other18 files · 14.0 MB
Repository1 file · 2.9 KB
Every file
FileTypeSizeSHA-256
transformer/diffusion_pytorch_model-00001-of-00002.safetensorsWeights5.0 GB 116134e5492c
transformer/diffusion_pytorch_model-00002-of-00002.safetensorsWeights1.7 GB 3b4c0aafe270
vae/diffusion_pytorch_model.safetensorsWeights1.4 GB 230496cb59ff
vision_encoder/model.safetensorsWeights978.7 MB 2180ad739ecc
assets/diffusers_outputs/edge_action_id_av_inverse_0_diffusers.jsonConfiguration15.9 KB
assets/diffusers_outputs/edge_action_id_av_inverse_1_diffusers.jsonConfiguration15.7 KB
assets/edge_action_id_av_0_output.jsonConfiguration11.2 KB
assets/edge_action_id_av_1_output.jsonConfiguration11.0 KB
assets/example_action_fd_umi_action_chunks.jsonConfiguration10.3 KB
assets/example_i2v_prompt.jsonConfiguration9.2 KB
assets/example_reasoning_prompt.jsonConfiguration151 B
assets/negative_prompt.jsonConfiguration17.6 KB
config.jsonConfiguration1.7 KB
generation_config.jsonConfiguration154 B
model.safetensors.index.jsonConfiguration68.6 KB
model_index.jsonConfiguration555 B
modular_model_index.jsonConfiguration1.4 KB
preprocessor_config.jsonConfiguration334 B
processor_config.jsonConfiguration875 B
scheduler/scheduler_config.jsonConfiguration889 B
special_tokens_map.jsonConfiguration563 B
text_tokenizer/special_tokens_map.jsonConfiguration563 B
transformer/config.jsonConfiguration1.1 KB
transformer/diffusion_pytorch_model.safetensors.index.jsonConfiguration53.6 KB
vae/config.jsonConfiguration1.8 KB
video_preprocessor_config.jsonConfiguration367 B
BIAS.mdDocumentation4.7 KB
EXPLAINABILITY.mdDocumentation3.2 KB
PRIVACY.mdDocumentation1.2 KB
README.mdDocumentation53.0 KB
SAFETY.mdDocumentation3.7 KB
assets/benchmark-image2video.pngOther58.5 KB
assets/benchmark-overall.pngOther81.0 KB fdc1dbbeda37
assets/diffusers_outputs/edge_action_fd_diffusers_chunk_00.mp4Other14.8 KB
assets/diffusers_outputs/edge_action_fd_diffusers_chunk_01.mp4Other13.2 KB
assets/diffusers_outputs/edge_action_fd_umi_2chunk_diffusers.mp4Other23.6 KB
assets/diffusers_outputs/edge_i2v_diffusers.mp4Other1.3 MB 2fe93251039d
assets/edge_action_fd_umi_2chunk_output.mp4Other32.8 KB 10037bf0c2a2
assets/edge_action_id_av_0_output.pngOther113.8 KB 18b90a143ee0
assets/edge_action_id_av_1_output.pngOther102.5 KB 660085cebabd
assets/edge_i2v_output.mp4Other8.1 MB d4e87cbc2efe
assets/example_action_fd_umi_first_frame.pngOther67.0 KB
assets/example_action_id_av_0_input.mp4Other1.1 MB ff205f86ae16
assets/example_action_id_av_1_input.mp4Other1.6 MB 169e65cee76e
assets/example_i2v_input.jpgOther860.8 KB 1de51eb5c6d5
assets/example_reasoning_input.pngOther230.7 KB 6686b937bdb2
chat_template.jinjaOther12.2 KB
images/benchmark-reasoning.pngOther265.8 KB 75974a959d0e
text_tokenizer/chat_template.jinjaOther12.2 KB
.gitattributesRepository2.9 KB
text_tokenizer/tokenizer.jsonTokenizer17.1 MB 4dc692a99dca
text_tokenizer/tokenizer_config.jsonTokenizer393 B
tokenizer.jsonTokenizer17.1 MB 4dc692a99dca
tokenizer_config.jsonTokenizer177.3 KB

License and Download

License
other
Access
Open weights, no gate
Download size
9.1 GB
Download from NVIDIA

Released by NVIDIA through its official repository on Hugging Face.

Memory Requirements

PrecisionWeights in memory
As published9.1 GB
16-bit7.7 GB
8-bit3.9 GB
4-bit1.9 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Questions About Cosmos3-Edge

How much GPU memory does Cosmos3-Edge need?

About 9.3 GB at 16-bit and 2.3 GB at 4-bit: the weights (3.9B parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run Cosmos3-Edge on?

At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.

What license is Cosmos3-Edge released under?

other, as its publisher declares it. Read the license text before commercial use.

What is Cosmos3-Edge's context length?

131,072 tokens, from the maximum position embeddings in its published configuration.