Google cuts AI video costs by 30% with Gemini Omni 1.2 Flash

Detailed summary

Google has introduced Gemini Omni 1.2 Flash, an updated video-generation model that can create clips of up to 15 seconds, maintain greater visual consistency and support…

Google has introduced Gemini Omni 1.2 Flash, an updated video-generation model that can create clips of up to 15 seconds, maintain greater visual consistency and support conversational editing.

The updated model allows users to generate videos ranging from three to 15 seconds in one-second increments, up from the previous 10-second limit. Google's Gemini Omni 1.2 Flash is now available through the Gemini API and Google Cloud, with 720p video generation priced at $0.07 per second, 30% lower than its predecessor.

One of the key changes is improved reference adherence. Users can combine character, product and style images with motion clips of up to three seconds and audio cues to guide video generation. The model is designed to maintain details such as facial features, clothing, product packaging, typography and brand colours across different camera angles and scenes.

Gemini Omni 1.2 Flash also introduces multi-turn conversational editing, allowing users to modify an existing video through follow-up instructions rather than generating an entirely new clip. Changes can include replacing products or backgrounds, adjusting lighting and weather, and changing aspect ratios.

Google says the model also delivers smoother dialogue, improved lip-syncing and more controlled camera movements, while reducing unwanted scene transitions and visual inconsistencies. Users can generate initial clips at 360p or 720p before upscaling selected outputs to 1080p or 4K.

Generated videos also carry an imperceptible SynthID watermark, according to Google.

Opening of the article, excerpted from The Daily Star. To read the rest, open the article on the publisher's site.