Model
Imagen 4 AI Image Generator
Imagen 4 is a Google text-to-image model with a compact, control-oriented set of options. Alongside the prompt it accepts a negative prompt and a seed, and it covers square, portrait and landscape ratios. It does not take reference images, so everything is described in words.
What Imagen 4 does well
Negative prompt support
Exclude unwanted elements with a negative prompt of up to 5,000 characters.
Seeds for revisiting results
A seed value lets you return to a result and adjust the prompt from the same starting point.
Simple format choice
Aspect ratios include 1:1, 16:9, 9:16, 3:4 and 4:3.
Prompt-only workflow
As a pure text-to-image model, it keeps the process centered on writing clear, descriptive prompts.
Best for
- Descriptive text-to-image prompts without references
- Cleaning up results with negative prompts
- Repeatable variations using seeds
- Landscape and portrait photography-style scenes
- Blog, article and social illustrations
Prompt examples
Alpine Lake at Dawn
photograph of a still alpine lake at dawn, mirror reflection of jagged peaks, a red canoe pulled up on the pebble shore, thin mist over the water, cool blue tones with the first pink light on the summits
Add a negative prompt such as 'people, buildings, text' to keep the scene untouched.
Try this promptKitchen Garden Illustration
watercolor-style illustration of a small kitchen garden in raised wooden beds, tomatoes, basil and marigolds, a watering can on the path, soft morning light, gentle pastel palette
Naming three specific plants gives the beds variety in shape and color without overcrowding them.
Try this promptRemote Work Article Header
conceptual image for an article about remote work, a laptop on a picnic blanket in a sunny meadow, a cup of coffee and a small notebook, shallow depth of field, bright natural colors, landscape composition
Stating the article topic up front helps the model choose props that support the story.
Try this prompt
Frequently asked questions
- Does Imagen 4 accept reference images?
- No. Imagen 4 is text-to-image only. If you need to guide output with images, use a model such as Nano Banana 2 or FLUX.2.
- Can I use negative prompts and seeds with Imagen 4?
- Yes, both are supported. The negative prompt can be up to 5,000 characters long.
- What sizes can Imagen 4 produce?
- You choose an aspect ratio: 1:1, 16:9, 9:16, 3:4 or 4:3. There is no separate resolution setting for this model.
Related models
Try Imagen 4
Or let Smart mode choose the best model for each request.
Start creatingCapabilities in OpenDream
| Capability | Imagen 4 |
|---|---|
| Text to image | Yes |
| Edit / image to image | No |
| Reference images | No |
| Resolutions | Automatic |
| Aspect ratios | 5 |
| Negative prompt | Yes |
| Seed | Yes |
| Transparent background | No |
| Speed | Medium |
| Quality | High |
| Credits per image (from) | 8 |
Live from the OpenDream model registry. Costs can vary with resolution and reference images; the exact cost is shown before you generate.