Google has reached the published shutdown date for three stable Imagen 4 image-generation models in the Gemini API.
According to the official Gemini deprecations page, the following models are shut down as of August 17, 2026:
| Model | Release Date | Shutdown Date | Recommended Replacement |
|---|---|---|---|
| imagen-4.0-generate-001 | June 24, 2025 | August 17, 2026 | gemini-3.1-flash-image |
| imagen-4.0-ultra-generate-001 | June 24, 2025 | August 17, 2026 | gemini-3.1-flash-image |
| imagen-4.0-fast-generate-001 | June 24, 2025 | August 17, 2026 | gemini-3.1-flash-image |
Google defines a “shutdown” as the point at which a model is completely turned off and the endpoint is no longer available. Deprecation announcements appear on the Release notes page; earliest shutdown dates are tracked on the deprecations page.
The official recommended replacement for all three Imagen 4 endpoints is gemini-3.1-flash-image (also referred to in related Google documentation as part of the “Nano Banana” Gemini image model family).
Primary source (copy-and-paste link):
https://ai.google.dev/gemini-api/docs/deprecations
Data notes for operators / HNWI / enterprise AI workflows
- The three models share the identical shutdown date and the same listed replacement.
- Related Firebase documentation confirms the August 17, 2026 earliest shutdown applies across the Gemini Developer API and Agent Platform (formerly Vertex AI) and advises migration to Gemini 3.x Image models before the date to avoid service disruption.
- gemini-3.1-flash-image image-output pricing (official Gemini API pricing page) is token-based at $60 per 1M image-output tokens, equating to approximately $0.045 (0.5K), $0.067 (1K), $0.101 (2K), and $0.151 (4K) per image depending on resolution.
This is a hard endpoint retirement, not a soft deprecation. Any production image-generation workloads still calling the three imagen-4.0-*-001 model IDs will fail once the shutdown is fully enforced.