Google has introduced Gemini Omni 1.1 Flash, a suite of generative video capabilities designed for developers building creative tools and media software. The model is now production-ready via the Gemini API and Google AI Studio, according to the company.
One major feature is scene extension, which allows developers to continue generating video footage seamlessly from where existing content ends. Omni 1.1 can analyze up to 10 seconds of prior video context—a significant improvement from previous models that only referenced the final second. This approach improves visual consistency and narrative continuity, and videos can be extended in 10-second increments up to 40 seconds total.
Another capability lets developers specify starting and ending frames to generate smooth transitions and camera movements between two keyframes. This feature is designed for complex camera orbits, zoom transitions, or seamless looping clips without jump cuts.
To accelerate prototyping workflows, Omni 1.1 can generate lightweight preview videos in 360p resolution up to 60% faster and at one-third the cost compared to standard 720p generation. For professional-grade output, the model can upscale videos to 1080p or 4K resolution ready for production use.
Developers can also reference up to three seconds of video when crafting scenes, allowing them to maintain visual context and character consistency based on video inputs. This multimodal capability enables more precise control over generated content.
Omni 1.1 is rolling out across Google’s developer ecosystem. It is immediately available in Google AI Studio and via the Gemini Enterprise Agent Platform API for enterprise customers. The model is also accessible to Google AI Plus, Pro, and Ultra subscribers in Google Flow globally, with scene extension available in the Gemini app for all subscription tiers.
Key facts
- Scene extension now analyzes up to 10 seconds of prior context versus only the final second in previous models
- Videos can be extended in 10-second increments up to 40 seconds total cumulative length
- 360p preview generation is up to 60% faster and costs one-third as much as 720p generation
- Developers can specify first and last frames to generate smooth camera transitions
- Models can upscale to 1080p or 4K for professional production
- Up to three seconds of video can be referenced in multimodal prompts