Skip to main content
AI image model
Google
Google Imagen

Google Imagen4

Google Imagen4 is Kyeo AI's text-only, single-image workflow at 8 credits per request. It accepts a main prompt of up to 5,000 characters, an optional negative prompt, five aspect ratios, and a text seed; the default ratio is 1:1. `Imagen4` is the upstream workflow name, not a quality guarantee.

Capabilities
Text-only image generation
Credit cost
8 credits per request
Prompt limit
5,000 characters
Image input
Not supported
The upstream interface route is `google/imagen4`; Kyeo sends text and supplied schema fields only.
The main prompt allows 5,000 characters, image input is unsupported, and the current output is one image.
The form offers a negative prompt, five ratios, and a text seed, with 1:1 as the default ratio.
Each request costs 8 credits.
30-second overview
Google Imagen
8 credits per request
What it does best
This is an upstream interface text-to-image route. Kyeo does not accept a reference image or expose editing, inpainting, upscaling, output count, resolution, format, prompt rewriting, or safety-threshold controls.
Best for
Use it to compare one text brief at 1:1, 16:9, 9:16, 3:4, and 4:3, and to record what the upstream interface's negative-prompt and text-seed fields change in practice.
Popular searches
how to use Google Imagen4 on KyeoGoogle Imagen4 credit costGoogle Imagen4 prompt limit

At a glance

What is this model like?

Kyeo sends the main prompt and any supplied `negative_prompt`, `aspect_ratio`, and text `seed` to `google/imagen4`; it does not send an image. The upstream interface still lists this workflow. Google's direct Imagen 4 Generate workflow is deprecated, and the upstream interface does not say whether it uses that model, another upstream service, or its own parameter conversion.

What it does best
This is an upstream interface text-to-image route. Kyeo does not accept a reference image or expose editing, inpainting, upscaling, output count, resolution, format, prompt rewriting, or safety-threshold controls.
Best for
Use it to compare one text brief at 1:1, 16:9, 9:16, 3:4, and 4:3, and to record what the upstream interface's negative-prompt and text-seed fields change in practice.
Why use it on Kyeo AI
Before submission, Kyeo shows the 5,000-character limit, five ratios, the 1:1 default, a text seed, and the fixed 8-credit charge.

Key facts

Quickly assess whether this model matches your workflow.

Category
AI image model
Vendor
Google
Model family
Google Imagen
Integration
Current site integration
Platform model ID
google/imagen4
Runtime
Asynchronous task
Prompt limit
5,000 characters
Image input
Not supported

Also known as

The same model may appear under different names across documentation and community discussions; this list keeps them easy to search and compare.

google/imagen4
Google Imagen4
Google Imagen 4

Common search intents

These questions and keywords combine public references with common creative use cases to help you choose faster.

how to use Google Imagen4 on Kyeo
Google Imagen4 credit cost
Google Imagen4 prompt limit
Google Imagen4 supported ratios
does Google Imagen4 accept images
Google Imagen4 negative prompt limit
Google Imagen4 seed limit
how many images does Google Imagen4 return
Google Imagen4 vs Imagen4 Fast
Which Imagen4 controls are available on Kyeo

Selection guide

Use these decision points when choosing a model.

1
Choose it for text-only generation with a 1:1 default

Use the workflow when no reference image is needed and one image in any of five fixed ratios is enough for the test.

2
Use another model for editing or output controls

This entry has no image input, output-count setting, size, format, or safety control. Choose a model that explicitly exposes the required control for source-image preservation, local edits, upscaling, or a specific file format.

3
Test both text-field limits

Upstream interface caps the negative prompt at 5,000 characters and seed at 500 characters, while Kyeo currently rejects neither overlong value. Hold the main prompt and ratio constant, then change one field at a time.

Popular comparisons

Common comparison questions help you find the practical differences quickly.

How do Kyeo's controls differ between Imagen4 and Imagen4 Fast?
How do Imagen4 and Imagen4 Fast compare on Kyeo credits?
Do Imagen4 and Ultra use different default ratios on Kyeo?
Which direct Imagen 4 controls are absent from Kyeo?
How should Imagen4 and Imagen4 Fast be tested with the same brief?
Are Imagen4 and 4o Image both text-only on Kyeo?

Model comparison

Compare the current model with alternatives at a glance.

Dimension
Google Imagen4
Google Imagen4 Fast
Google Imagen4 Ultra
Kyeo model ID
google/imagen4
google/imagen4-fast
google/imagen4-ultra
Kyeo input and output
Text only, fixed single-image output
Text only, fixed single-image output
Text only, fixed single-image output
Default aspect ratio
1:1
16:9
1:1
Kyeo seed input type
Text, no enforced length limit
Number, minimum 0
Text, no enforced length limit
Kyeo credits
Fixed 8
Fixed 4
Fixed 12

Practical usage insights

Practical observations based on real creative workflows.

Google Imagen: Fast to explore

The upstream interface uses `google/imagen4` but does not say whether it maps to `imagen-4.0-generate-001`, another Google image workflow, or a conversion layer.

Google Imagen: Useful for comparisons

On March 24, 2026, Google deprecated `imagen-4.0-generate-001` and recommended migration to `gemini-2.5-flash-image` before June 30.

Capabilities

Google Imagen: Clear input options

Kyeo places the main brief in `prompt` and forwards a supplied `negative_prompt`, `aspect_ratio`, and text `seed`. It accepts no reference image and uses only `google/imagen4`.

Google Imagen: Practical controls

Kyeo exposes the five fixed ratios. Kyeo currently enforces only the main-prompt limit. Technical identifiers: `auto`.

Google Imagen: Transparent credit use

Google's direct Generate page says negative prompts are unsupported, so field effects, seed repeatability, output format, watermarking, and safety handling require the upstream interface result test.

Use cases

Ratio comparison from a 1:1 baseline

Start with the default square output, then use the same brief at landscape, portrait, 3:4, and 4:3. Check subject placement, cropping, text, and negative space against the destination layout.

Negative-prompt and seed comparison

Hold the main prompt and ratio constant while changing `negative_prompt` and text `seed` separately.

Google Imagen: Option comparison

Use the same main prompt, ratio, and exclusions in Fast, Imagen4, and Ultra, then record credits, elapsed time, output differences, and failures. Names and prices are not substitutes for test results.

Prompt tips

Describe visible content in the main prompt

Cover the subject, action, setting, composition, lighting, and required text.

Keep both text parameters within upstream interface limits

Limit the negative prompt to 5,000 characters and the text seed to 500 characters. Kyeo does not yet reject either overlong value before submission.

Change one variable in seed tests

Fix the main prompt, negative prompt, and ratio; repeat the same seed; then change only the seed. Describe repeatability only after the current upstream interface route produces consistent results.

Strengths

The text-only input, 5,000-character main-prompt limit, and fixed single-image output are explicit.
Five ratios, a 1:1 default, and a text seed are visible before submission.
Each request costs 8 credits regardless of the exposed parameter values.
The upstream interface schema can be checked field by field against what Kyeo forwards.

Limitations

Google deprecated `imagen-4.0-generate-001` and recommended migration.
Kyeo accepts text only and does not expose image editing, output count, resolution, format, safety settings, or prompt rewriting.
Upstream interface caps the negative prompt at 5,000 characters and seed at 500 characters, but Kyeo's schema lacks both limits; overlong values may fail only after reaching the upstream service.
The upstream interface exposes negative prompt and text seed, while Google's direct Generate page says negative prompts are unsupported.

Adjustable parameters

Quickly assess whether this model matches your workflow.

Negative prompt
negative_prompt
Optional
Parameter type: text
Aspect ratio
aspect_ratio
Optional
Parameter type: select
Default: 1:1
1:1
16:9
9:16
3:4
4:3
Seed
seed
Optional
Parameter type: text
not enforced by Kyeo

Credit usage

8 credits per request

Kyeo charges 8 credits for each single-image Google Imagen4 request.

Budget tip

Before batch generation, run an A/B test with the same assets across the current model and alternatives to validate quality and avoid wasting credits.

FAQ

Related models

Compare these similar candidates before deciding.

Google Imagen4 Fast

Google Imagen4 Fast is Kyeo AI's text-only, single-image workflow at 4 credits per request. It accepts a main prompt of up to 5,000 characters, an optional negative prompt, five aspect ratios, and a seed; the default ratio is 16:9. `Fast` is a workflow name, not a measured promise about latency or quality.

AI image model
4 credits per request

Google Imagen4 Ultra

Google Imagen4 Ultra is Kyeo AI's text-only, single-image workflow at 12 credits per request. It accepts a main prompt of up to 5,000 characters, an optional negative prompt, five aspect ratios, and a text seed. `Ultra` is a workflow name, not a Kyeo guarantee that it outperforms the other Imagen4 routes.

AI image model
12 credits per request

4o Image

4o Image is Kyeo AI's text-to-image implementation of the upstream `gpt4o image generation` workflow. It offers `1:1 / 3:2 / 2:3`, costs a fixed 3 credits, and accepts no reference images.

AI image model
3 credits / request

Sources

Content is based on model documentation, feature references, and the settings available on Kyeo AI.

Source notes

This page treats Kyeo's current configuration, request path, and the upstream interface schema as the available scope. The entry accepts text only, uses `google/imagen4`, limits the main prompt to 5,000 characters, forwards `negative_prompt`, `aspect_ratio`, and a text `seed` when supplied, produces the currently configured single image, and costs 8 credits. The upstream interface also offers an `auto` ratio, caps the negative prompt at 5,000 characters, and caps seed at 500 characters. Kyeo shows five fixed ratios with a 1:1 default but does not yet enforce either text-field limit. Google's direct `imagen-4.0-generate-001` documentation says negative prompts are unsupported, and Google deprecated that workflow on March 24, 2026, with migration recommended before June 30.

Last reviewed: 2026-07-20
Google Cloud: Imagen 4 model documentation
Official

Google describes `imagen-4.0-generate-001` as text-to-image with five ratios, PNG or JPEG, up to four direct outputs, and 1K or 2K resolutions, while explicitly listing negative prompts and image editing as unsupported.

View source
Google Cloud: Vertex AI release notes
Official

Used for Google Imagen lifecycle context; Kyeo's current options remain those shown on the page and in the workbench.

View source
Google Cloud: Generate images from text
Official

Google's direct guide establishes Imagen 4 family input, output, and lifecycle context.

View source
Upstream interface documentation: Google Imagen4
Platform

The upstream schema documents a 5,000-character main prompt and negative prompt, six aspect ratios including `auto`, and a text seed of up to 500 characters; the example workflow is `google/imagen4`.

Kyeo AI model configuration
Platform

Kyeo currently accepts text only, limits the main prompt to 5,000 characters, offers five fixed ratios with a 1:1 default, uses a text seed, returns the configured single image, and charges 8 credits.