Imagen 4 - Kyma API

Overview

Imagen 4 is a text-to-image model from Google that produces photorealistic humans, sharp embedded text, and richly composed scenes. It operates as the default quality tier, accepting text prompts and returning image outputs. On Kyma, the model runs through an OpenAI-compatible endpoint with automatic failover across serving paths. Every request reports exact cost in usage.cost and the active model via the X-Kyma-Model header. Prompt caching is enabled, billing repeated prompt prefixes at this model’s cached input rate. The model does not support reasoning, vision input, or structured outputs. It generates images on a per-image basis at a medium speed tier and is positioned at a balanced cost tier.

Specs

Field	Value
Model ID	`imagen-4`
Best for	Photoreal portraits, branded scenes, hero images
Context window	0
Input modalities	Text
Output modalities	Image
Tool calling	Yes
Structured outputs	Yes
Prompt caching	Yes
Speed	Medium
Cost band	Balanced
Release stage	Stable

Pricing

	Per 1M tokens
Input	$0.20
Output	$0.60

Use this when

Photorealistic Portrait Generation — Generate lifelike human portraits with accurate lighting and skin tones.
Branded Marketing Visuals — Produce campaign-ready images with sharp, legible embedded text.
Website Hero Imagery — Create richly composed header graphics for landing pages.
Product Scene Mockups — Render realistic product placements in controlled environments.

Not ideal for

Do not use this model for tasks requiring image-to-image editing, structured JSON outputs, or sub-second generation latency.

Pick something else when

You need faster generation latency: use imagen-4-fast.
You require maximum photorealism and detail: use imagen-4-ultra.
You need scalable vector graphics: use recraft-v4-vector.

Example

from openai import OpenAI

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

response = client.chat.completions.create(
    model="imagen-4",
    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 Imagen 4 support image-to-image or vision inputs? No. The model only accepts text prompts and outputs images; it does not process visual inputs. How does prompt caching work for this model? Kyma caches repeated prompt prefixes and bills them at this model’s cached input rate, reducing costs for iterative prompting. Can I get structured outputs like JSON from this model? No. Imagen 4 does not support structured outputs or reasoning capabilities.

​Overview

​Specs

​Pricing

​Use this when

​Not ideal for

​Pick something else when

​Example

​Agent query example

​FAQ