ZMIME
NVIDIA · image model · Release date unconfirmed

Cosmos3-Super-Text2Image (Agentic)

Cosmos3-Super-Text2Image (Agentic) is an image model from NVIDIA. It handles text to image.

  • Image model
  • Open weights
  • Text to image
Arena score
1,164#27 of 82 · text to image
Starting price
—Not published
Largest size
—Not published
References
—Text prompts only
Quality & price

How it rates and what it costs

Text to image arena
1,164 ±7#27 of 82 · 10,803 votes

Arena scores come from blind, side-by-side human votes on the same prompts (higher is better; ± is the 95% range). Boards updated Oct 26, 2026.

Specifications

The details

Lab
NVIDIA
Type
Image model
Released
—
Inputs
Text prompts
Output
Images
Can do
Text to image
Weights
Open weights · OpenMDW-1.1
Availability
Not on the API hosts we track

Arena listing

Same lab

More image models from NVIDIA

Questions

About Cosmos3-Super-Text2Image (Agentic)

How good is Cosmos3-Super-Text2Image (Agentic)?

In blind side-by-side votes, Cosmos3-Super-Text2Image (Agentic) scores 1,164 for text to image (#27 of 82). Higher is better; scores within about 14 points of each other are effectively tied.

Who makes Cosmos3-Super-Text2Image (Agentic)?

Cosmos3-Super-Text2Image (Agentic) is an image model from NVIDIA.

What can Cosmos3-Super-Text2Image (Agentic) do?

Cosmos3-Super-Text2Image (Agentic) accepts text prompts and produces images. In practice: text to image.

Is Cosmos3-Super-Text2Image (Agentic) open source?

Yes. NVIDIA publishes the weights under the OpenMDW-1.1, so you can run it on your own hardware.

Where can I use Cosmos3-Super-Text2Image (Agentic)?

None of the API hosts we track list it yet.