Imagen 4 Ultra - Kyma API

Overview

Imagen 4 Ultra is Google’s highest-fidelity text-to-image model. It generates images with fine-grained detail, accurate lighting, and high resolution suitable for physical print and large-format displays. On Kyma, the model accepts text prompts and returns image outputs. It operates on a premium cost tier, bills per image, and has the slowest generation speed in the lineup. Prompt caching is enabled, billing repeated prompt prefixes at this model’s cached input rate. Every request includes automatic failover and returns exact usage costs in the response payload. The model does not support image inputs, reasoning, vision, or structured outputs. Generation latency is intentionally higher to prioritize quality, making it unsuitable for real-time or high-throughput workflows.

Specs

Field	Value
Model ID	`imagen-4-ultra`
Best for	Print-ready hero assets, large displays, premium campaigns
Context window	0
Input modalities	Text
Output modalities	Image
Tool calling	Yes
Structured outputs	Yes
Prompt caching	Yes
Speed	Slow
Cost band	Premium
Release stage	Stable

Pricing

	Per 1M tokens
Input	$0.20
Output	$0.60

Use this when

Print-ready marketing assets — Generate high-resolution visuals that maintain sharpness and color accuracy at physical print sizes.
Hero banner generation — Create detailed background images for website headers and premium digital campaigns.
High-resolution product mockups — Produce detailed visual assets requiring precise texture, lighting, and material representation.
Archival quality illustrations — Render long-form creative projects where maximum detail is prioritized over generation speed.

Not ideal for

Real-time interactive applications, rapid prototyping, or high-volume batch generation where low latency and cost efficiency are primary constraints.

Pick something else when

You need faster generation for rapid iteration: use imagen-4-fast or imagen-4.
You need vector graphics or scalable layouts: use recraft-v4-vector or recraft-v4-vector-pro.
You need lower-cost bulk generation: use flux-2-pro or nano-banana.
You need image-to-image editing: use recraft-v4 or flux-kontext-pro.

Example

from openai import OpenAI

client = OpenAI(base_url="https://kymaapi.com/v1", api_key="ky-...")

response = client.chat.completions.create(
    model="imagen-4-ultra",
    messages=[{"role": "user", "content": "..."}],
)

Agent query example

Ask the API which models fit, instead of hardcoding an id:

curl "https://kymaapi.com/v1/models?recommended_for=generate-images&tools=true&quality_tier=frontier-open"

FAQ

Does this model support image-to-image or reference inputs? No, it only accepts text prompts and outputs images. How does prompt caching work for this model? Repeated prompt prefixes are billed at this model’s cached input rate, reducing costs for workflows with consistent system prompts. What happens if the serving path degrades during generation? Kyma automatically reroutes the request to a healthy path via built-in failover, so you do not need to handle retries in your code.

​Overview

​Specs

​Pricing

​Use this when

​Not ideal for

​Pick something else when

​Example

​Agent query example

​FAQ