Prompt expansion
When you turn on Enable Prompt Expansion, your prompt is rewritten in more detail before the image is made. Your own prompt is used if the rewrite fails.
Create a picture from a description
Cosmos 3 Super · Text→Image
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.
Made by NVIDIA400 creditsper run Pro
Opens the image studio with Cosmos 3 Super · Text→Image ready to use. You can look around as a guest. Running it takes the Pro plan or higher.
What Cosmos 3 Super · Text→Image makes, what it needs and what you get
What it makes
Create a picture from a description
Turns a written description into a picture.
What you need
- A description of the picture
What you get
A new picture
- JPEG and PNG files
About Cosmos 3 Super · Text→Image
Cosmos 3 Super is an Nvidia model that makes an image from a written prompt. You describe the picture and get an image back.
A text prompt is the only required input. When you turn on Enable Prompt Expansion, your prompt is rewritten into a more detailed, structured form before the image is made. If the rewrite fails, your own prompt is used. When you turn on Enable Agentic Generation, the model makes several candidate images, compares them and refines the prompt between rounds. This can match your request better, but it takes longer and increases the number of image generations you are billed for. Both switches start off.
With agentic generation on, Agentic Max Iterations (1 to 3) sets the refinement rounds. Agentic Samples Per Iteration (1 to 3) sets the candidates per round. Agentic Early Stop stops the search early when a candidate is already a strong match, and it starts on. Image Size, a step count (1 to 50) and Guidance scale (0 to 20) also adjust the result. Cosmos 3 Super costs 400 credits.
Other models for this job
Same job: create a picture from a description. Compare before you choose.
Image Free Open
SDXL LightningCreate a picture from a description
A speed-tuned variant of Stable Diffusion XL that generates high-quality images in a handful of steps, trading a little fine detail for near-instant results - ideal for rapid concept exploration and draft iteration.
ByteDance10 creditsper run
Image Free Open
Qwen Image LightningCreate a picture from a description
Alibaba's fast-inference Qwen Image variant, tuned for near-instant text-to-image generation without sacrificing prompt accuracy - built for rapid drafts and high-volume workflows.
Alibaba10 creditsper run
Image Free Open
Qwen ImageCreate a picture from a description
Alibaba's Qwen Image model generates accurate, detailed visuals from text prompts with strong prompt comprehension, including legible in-image text and complex compositions.
Alibaba10 creditsper run
Image Free Open
Flux KreaCreate a picture from a description
A FLUX variant tuned by Krea for stylized, painterly image generation - strong for concept art and creative visuals that want more artistic flair than strict photorealism.
Black Forest Labs85 creditsper run
What it does well
Agentic generation
When you turn on Enable Agentic Generation, the model makes and compares several candidate images, then refines the prompt between rounds to match your request better.
Control over refinement
Allow 1 to 3 rounds with 1 to 3 candidates each. Agentic Early Stop stops early when a candidate is already a strong match.
Step count and quality
The step count runs from 1 to 50 and starts at 28. More steps give higher quality but take longer.
How to use Cosmos 3 Super
How It Works
How It Works
01
Enter your prompt
Enter a description of the image you want.
02
Choose the extras
Switch on Enable Prompt Expansion or Enable Agentic Generation if you want them. Adjust the other options if you wish.
03
Run Cosmos 3 Super
Run Cosmos 3 Super for 400 credits and receive your image.
Tip Image Size keeps each edge between 512 and 1280 pixels, in multiples of 16.
Good for
Illustrations from a written idea
Describe a scene in your own words and get an image back.
Concept art from a description
Describe a character or a setting and get an image back.
Questions
Quick answers to help you get the most out of Cosmos 3 Super · Text→Image in FunAI Studio.
1. What is required to start?
2. What does Enable Agentic Generation do?
3. How much does Cosmos 3 Super cost?