gen‑ai.news
ClosedFreemiumDiscontinued

Imagen

Google DeepMind

Imagen is an image generation model line from Google DeepMind. The latest release is Imagen 4 (20 May 2025). This page tracks all 4 Imagen releases since 23 May 2022, with specs, licensing, and how to access each one. Official site →

Imagen 4released

ClosedFreemiumOfficial APIConsumer app

Announced at Google I/O 2025; sharper fine details, text, and typography; faster generation

License
Proprietary
Commercial use
Allowed
Availability
Official API, Consumer app

Announced Google I/O May 20, 2025

Imagen 3released

ClosedFreemiumOfficial APIConsumer app

Higher detail and fewer artifacts than Imagen 2; improved natural lighting; available in Gemini

License
Proprietary
Commercial use
Allowed
Availability
Official API, Consumer app

Imagen 2released

ClosedFreemiumOfficial APIConsumer app

First public commercial availability via Vertex AI and Gemini; improved text rendering and photorealism

License
Proprietary
Commercial use
Allowed
Availability
Official API, Consumer app

Imagen 1released

Closed

Cascaded diffusion; photorealism at the time; no public access - research demo only

License
Proprietary
Commercial use
Restricted

FAQ

When was Imagen 4 released?
Imagen 4 was released on 20 May 2025.
Is Imagen open source?
No. Imagen 4 is closed: the weights are not released and it is available through an API or app only.
Can Imagen be used commercially?
Yes. Commercial use of Imagen 4 is allowed under the Proprietary licence.
Where can I use Imagen?
Imagen 4 is available via Official API and Consumer app.

See Imagen in context on the image model timeline, or browse all model families.