Overview
Gemini Omni is the first full-modal generative model released by Google at the I/O conference on 2026-05-19, serving as the 'creation engine' within the Gemini family. It can generate any output (text, image, video, audio) from any input (text, image, video, audio), with all generated content automatically carrying SynthID watermarking. It is currently the single model with the most comprehensive input-output dimensions on the market.
On 2026-08-27, Google brought Gemini Omni 1.1 Flash (gemini-omni-1.1-flash) to general availability, a production-focused generative video model. It adds scene extension that reads up to 10 seconds of prior footage and extends in 10-second steps to a cumulative 40 seconds, first/last-frame interpolation for camera moves, a 3-second reference clip for character consistency, a 360p draft mode (~60% faster and a third of the cost of 720p), and upscaled 1080p/4K output. The old preview endpoint sunsets on 2026-10-09.
Key Features
- Full-Modal Generation: Text→Image/Video/Audio, Image→Video/Text, Video→Editing/Dubbing, any combination
- Built-in SynthID Watermark: All AI-generated content automatically embeds invisible watermarks for traceable origin
- Free YouTube Shorts Remix Trial: Use video generation capabilities via YouTube Shorts without subscription
- Tiered Subscription Access: AI Plus/Pro/Ultra subscribers can fully use it in the Gemini App
- World Understanding Capability: Google calls it a 'leap in world understanding', capable of simulating visual consistency of the physical world
- Gemini Omni 1.1 Flash (GA since 2026-08-27): Production-focused generative video model (gemini-omni-1.1-flash) with scene extension reading up to 10 seconds of prior footage and extending in 10-second steps to a cumulative 40 seconds
- First/Last-Frame Interpolation: Supports camera moves by interpolating between a first and last frame
- Reference-Clip Character Consistency: Uses a 3-second reference clip to keep characters consistent across shots
- 360p Draft Mode: Draft mode is ~60% faster and a third of the cost of 720p
- Upscaled 1080p/4K Output: Output can be upscaled to 1080p and 4K
Use Cases
- Short video creation (YouTube Shorts/TikTok content)
- One-stop multimodal content production
- Image/video style transfer
- Visual explanation generation for educational scenarios
- Scene extension for longer video sequences
- Camera-move generation via first/last-frame interpolation
- Character-consistent multi-shot video production
Pros
- Highest freedom of full-modal combination (any input→any output)
- Built-in SynthID watermark for worry-free compliance
- Free YouTube Shorts entry lowers the experience barrier
- Google infrastructure ensures high availability and global distribution
- Gemini Omni 1.1 Flash is GA as of 2026-08-27 with scene extension, first/last-frame interpolation, and reference-clip character consistency
- 360p draft mode is ~60% faster and a third of the cost of 720p, with upscaled 1080p/4K output
Pricing
YouTube Shorts Remix is free to use. Access in the Gemini App requires AI Plus ($20/mo), Pro ($50/mo), or Ultra ($200/mo) subscription. API pricing has not yet been announced.
Summary
Gemini Omni represents the first implementation of Google's 'modal freedom' vision, lowering the experience barrier through a free entry point via YouTube. With the 2026-08-27 general availability of Gemini Omni 1.1 Flash, the lineup now covers production-focused video generation with scene extension, first/last-frame interpolation, reference-clip character consistency, a fast 360p draft mode, and upscaled 1080p/4K output. It is suitable for short video creators and teams needing rapid generation of diverse content formats, while API-level integration awaits pricing announcement.
Version History
- Gemini Omni 1.1 Flash goes GA (2026-08-27): Google launched Gemini Omni 1.1 Flash (gemini-omni-1.1-flash), a production-focused generative video model: scene extension that reads up to 10 seconds of prior footage and extends in 10-second steps to a cumulative 40 seconds, first/last-frame interpolation for camera moves, a 3-second reference clip for character consistency, a 360p draft mode (~60% faster and a third of the cost of 720p), and upscaled 1080p/4K output. The old preview endpoint sunsets Sept 30.