Google releases Gemini 3.1 Flash Image and 3 Pro Image as GA with exclusive video-to-image generation

Google moved gemini-3.1-flash-image and gemini-3-pro-image to general availability on May 28, 2026, bringing native visual generation to the Gemini API's stable tier. The flash model gains an exclusive capability at launch: video-to-image generation, which allows developers to pass video files or YouTube URLs directly to generate thumbnails, movie posters, or infographics.

What's new

Models released to GA

• gemini-3.1-flash-image — the faster, lighter image-generation model. Previously available as gemini-3.1-flash-image-preview, now deprecated with a June 25, 2026 shutdown date.
• gemini-3-pro-image — the higher-capability image model. Previously available as gemini-3-pro-image-preview, also deprecated with a June 25 shutdown.

Video-to-image generation (exclusive to gemini-3.1-flash-image)

• Input: video files or YouTube URLs
• Output: images derived from the video content — thumbnails, movie posters, or infographics
• This capability is not available on the pro image model

Migration deadline: Developers using preview model IDs must migrate before June 25, 2026. The preview variants (-preview suffix) will be shut down at that date.

Context

The Gemini Image GA arrives alongside a broader Gemini 3.x family consolidation. On June 1, 2026, Google shut down the entire Gemini 2.0 Flash family — gemini-2.0-flash, gemini-2.0-flash-001, gemini-2.0-flash-lite, and gemini-2.0-flash-lite-001 — directing developers to Gemini 3.5 Flash or 3.1 Flash Lite as replacements. The 3.x generation is now the stable API tier across both text and image modalities.

These image model GAs follow the May 19, 2026 launch of Gemini 3.5 Flash (described by Google as "most intelligent model for sustained frontier performance on agentic and coding tasks") and the public preview of Managed Agents in Google-hosted Linux sandboxes. Together they represent the completion of the core 3.x API surface.

Why it matters

The video-to-image capability is the most novel element at GA. Converting a video or YouTube URL into a usable image is a content workflow that currently requires either manual screenshot selection or a separate pipeline step. Integrating it into the Gemini API means developers can automate that step within an existing image-generation workflow — passing a video file and receiving a thumbnail or poster back in the same API call.

From a developer planning perspective, the move from preview to GA matters for production commitments. GA models carry SLA guarantees and stability promises that preview variants do not. The June 25 shutdown creates a four-week migration window — short enough to act on now for any production usage of the old model IDs.