source: Google DeepMind: Gemini Omni 1.1 Flash lets you build with more control
level: technical
google deepmind launched gemini omni 1.1 flash, an update to its generative video model aimed at developers. the release adds scene extension, allowing users to continue existing footage using up to 10 seconds of prior context, a jump from the previous one second. videos can be extended in 10-second increments up to 40 seconds total. the model also supports specifying first and last frames to generate smooth transitions and camera movements between them.
the update introduces 360p preview generation, which is up to 60% faster and costs one third of standard 720p output, useful for rapid prototyping. final videos can be upscaled to 1080p or 4k resolution. developers can also include up to three seconds of video references in prompts to maintain visual consistency. the model is available through google ai studio and the gemini enterprise agent platform, with integrations already in adobe firefly, figma weave, and runway.
early adopters report strong accuracy and reliability. figma weave's creative director noted the model helps teams move beyond generating videos to directing them, while gmi cloud highlighted its value for educational content where details must be correct. pricing details were not fully disclosed, but the 360p tier is positioned as a cost-saving option. the release reflects a broader push to make generative video practical for professional workflows, not just experimental demos.
why it matters: developers can now build video tools with faster iteration and lower costs, making ai video production more accessible for real applications.
source: Google DeepMind: Gemini Omni 1.1 Flash lets you build with more control