Summary
Google says Gemini Omni 1.2 Flash can generate and edit videos up to 15 seconds long, with conversational edits and improved adherence to visual and audio references. The company has cut the price for 720p video by 30% from version 1.1, to $0.07 per second.
Google has launched Gemini Omni 1.2 Flash, a video-generation model that can create clips up to 15 seconds long and revise them through follow-up prompts. The company says the update also improves consistency when a video is guided by reference images, motion clips or audio cues. For 720p output, Google lists a price of $0.07 per second, 30% below Gemini Omni 1.1.
Google says the model is generally available through the Gemini API and Google Cloud, with access also rolling out across its apps and services.
Longer clips and reference-guided generation
Omni 1.2 Flash can generate or edit video at durations from 3 to 15 seconds, selectable in one-second increments. The previous version supported clips up to 10 seconds. Google presents the longer duration as a way to make a complete short sequence, such as a product explainer or social advertisement, in one generation.
The model can take combinations of character, product or style images, motion references up to three seconds long, and audio cues. Google says it is better at retaining details such as a character’s appearance, clothing, product packaging and brand colours across camera angles and transitions. The company’s examples include virtual try-on sequences, animated product displays and scenes built around reference images.
The update also supports editing a generated clip through successive conversational prompts. A user can, for example, request a change to the background, lighting or aspect ratio without restarting the video-generation process. Google says these edits work across the full 15-second clip and are intended to preserve other elements of the scene as changes are made.
Price, workflow and availability
The listed 720p rate is $0.07 per second, a 30% reduction from Omni 1.1. Google describes a workflow in which creators draft at 360p or 720p, make targeted revisions, then upscale selected clips to 1080p or 4K. The announcement does not give separate prices for those higher-resolution outputs.
Google also highlights more natural-sounding dialogue, clearer lip-sync and distinct voice conditioning for multiple speakers, alongside greater control over continuous camera movements. These are company-described capabilities; Google says development is continuing on fast sports action, rapid script changes and complex physics.
The model is available through the Gemini API and Google Cloud, and Google says Provisioned Throughput is available. In its apps, 15-second video generation in Google Vids is offered at no additional cost with a Google account or Google Workspace subscription. Access in the Gemini app is rolling out to Google AI Ultra subscribers; the company also lists Flow and Google Ads among the destinations for the model. Google says generated outputs carry an imperceptible SynthID watermark.