Imagen 4 Ultra

Google Imagen 4 highest fidelity — best detail, slowest gen. Print-ready hero assets at $0.081/image. Sync API.

Overview

imagen-4-ultra is Google’s top-fidelity Imagen 4 tier — highest detail, slowest generation. Use for print-ready hero assets, large displays, premium campaigns where fidelity beats cost.

Lives in the Imagen 4 family with imagen-4 (default) and imagen-4-fast (drafts). Same style language.

Specs

Field	Value
Model ID	imagen-4-ultra
Creator	Google
Backend	Vertex AI (imagen-4.0-ultra-generate-001, us-central1)
Best for	Print-ready hero assets, large displays, premium campaigns
Aspect ratios	1:1 (default), 3:4, 4:3, 16:9, 9:16
Sample count n	1–4
Pricing mode	Flat per image
Default latency	~20–40s
Output	Blob-hosted URL (Vercel CDN)

Pricing

Variant	Provider cost	Kyma list
imagen-4-ultra	$0.060	$0.081

Live source: GET https://kymaapi.com/v1/pricing.

Compared to other image models on Kyma

Strength	imagen-4-ultra	imagen-4	flux-2-pro	recraft-v4-pro
Top photoreal	★★★★★	★★★★	★★★★★	★★★★
Print fidelity	★★★★★	★★★	★★★★	★★★★★
Speed	★★★	★★★★	★★★★	★★★★
Cost	★★	★★★★	★★★★	★★

Use this when

• The deliverable is print-ready (poster, billboard, packaging, magazine).
• The image is a hero asset with downstream multi-format reuse.
• Fidelity is non-negotiable.

Pick something else when

• Volume/cost matters more than fidelity → imagen-4 or imagen-4-fast.
• You need a different style language (illustration, vector) → see Image generation.
• You need text-in-image legibility → gpt-image-2.

Example

curl -X POST https://kymaapi.com/v1/images/generations \
  -H "Authorization: Bearer $KYMA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "imagen-4-ultra",
    "prompt": "A cinematic editorial fashion portrait, soft natural light, 35mm film grain, magazine cover quality",
    "size": "9:16",
    "n": 1
  }'

Async endpoint — poll GET /v1/jobs/{id} until succeeded.