Google deprecates six video and image generation models in Gemini API: Veo 2/3 sunset June 30, Imagen 4.0 sunset August 17

Google announced on June 15, 2026 that six production video and image generation models in the Gemini API are being retired, with the first group -- three Veo models -- shutting down on June 30, 2026. Developers using these models have fewer than two weeks to migrate before API calls begin returning errors.

What's new

The Gemini API changelog published two deprecation notices on June 15:

Video generation (sunset June 30, 2026):

• veo-2.0-generate-001
• veo-3.0-generate-001
• veo-3.0-fast-generate-001

Migration path: move to veo-3.1-generate-preview or veo-3.1-fast-generate-preview, or the Veo 3.1 GA models available through the Gemini Enterprise Agent Platform.

Image generation (sunset August 17, 2026):

• imagen-4.0-generate-001
• imagen-4.0-ultra-generate-001
• imagen-4.0-fast-generate-001

Migration path: migrate to newer stable or preview endpoints via the Gemini deprecations page.

A third deprecation in the same update -- the experimental GMP Contextual View tool for grounding with Google Maps outputs -- took effect immediately on June 15.

Context

Veo and Imagen are Google's video and image generation APIs respectively, competing with Runway, Suno, and Midjourney in the generative media space. Both have gone through rapid generational updates over the last 18 months.

Veo 3.0 models were released as the production-grade video generation tier. The 3.1 replacements were introduced in preview, and the full migration cadence compresses existing models' lifecycles. Google has been advancing these families quickly -- the gap between Veo 3.0 release and its sunset is shorter than the comparable cycle for the 2.0 generation.

The Imagen 4.0 deprecation follows the same pattern. The August 17 deadline gives developers roughly two months to migrate, longer than the Veo window but still aggressive for production workloads that may have hardcoded model IDs.

Why it matters

For developers and production teams, the Veo June 30 deadline is the immediate concern. Any application calling veo-2.0-generate-001, veo-3.0-generate-001, or veo-3.0-fast-generate-001 will stop working in under two weeks. Unlike some API deprecations that include a grace period or soft failure mode, sunset dates in the Gemini API typically mean hard 4xx errors on the first day after cutoff.

The rapid succession of model retirements across Veo and Imagen reflects the pace of development in generative video and image -- a generation of capability now arrives roughly every six to nine months. The business implication for enterprises building on these APIs is a need for flexible model IDs (avoiding hardcoded versions) and active deprecation monitoring, since the windows between announcement and shutdown have shortened.