Gemini Omni is Google's multimodal video model line, and its defining trait is that audio is generated in the same pass as the picture rather than added afterward.
Omni Flash is the variant developers and platforms actually use. Google brought it to the Gemini API and AI Studio on 2026-06-30, and it is the second most-used generation tool on the 8frame canvas: 70 generations in the last 45 days, at an average of 65 credits each. Google prices video output on its own API at $17.50 per 1M output tokens, which its pricing page puts at roughly $0.10 per second of 720p video. On 8frame an 8-second clip with audio is 135 credits, about $1.35 at pack rate.
How Gemini Omni works
The model takes text, images, and video as conditioning and returns a clip with synchronized audio. It reasons about what the scene should sound like from what it renders, so footsteps land on footfalls and a room's ambience matches the room. You do not get an audio toggle on the canvas, because there isn't one in the model's default behavior: every Omni clip arrives with sound.
Google's Omni Flash documentation lists the model as gemini-omni-1.1-flash with 16:9 (default) and 9:16 aspect ratios, 360p and 720p (default) native resolutions, and 1080p and 4K available as upscales rather than native renders. Uploaded video for editing is capped at 10 seconds, and clips can be extended to 40 seconds total across turns. One regional catch worth knowing: editing or extending an uploaded video is not available in the European Economic Area, Switzerland, or the UK.
Google announced the 1.1 generation on 2026-08-27, adding scene extension, keyframe control, and up to 10 seconds of prior context when extending, against earlier models that referenced only the final frame.
What it costs where
| Route | Price | Notes |
|---|---|---|
| Gemini API | About $0.10 per second at 720p, billed as $17.50 per 1M output tokens | Google's published rate |
| Google AI subscription | Consumer tiers with monthly generation allowances | Access through the Gemini app and Flow |
| 8frame canvas | 51 credits at 3s, 85 at 5s, 135 at 8s, 169 at 10s | Audio always included, no watermark |
Credits are $0.01 each in a top-up pack, and the $19 Starter plan carries 1,000 credits a month, so about 7 full 8-second Omni clips or 19 three-second ones.
Examples
Drafting with sound in the loop. Because audio is not a separate step, an Omni draft tells you whether a shot works as a shot, not just as a picture. Three-second Omni passes at 51 credits are the cheapest way we have found to check whether a beat lands before committing to an expensive render.
Reference-driven consistency. On 8frame the tool auto-selects its mode from what you put on the canvas: prompt alone runs text-to-video, an attached still runs image-to-video, one to seven reference images run reference-to-video, and an attached clip runs video-edit. That last mode prices off the input clip's length, clamped to the 3 to 10 second window.
Related concepts
Gemini Omni vs Veo 3.1 puts the two Google video models side by side on price and output spec, which is the comparison most people actually need. What is Veo 3.1 Lite covers the budget tier that competes with Omni on cost, and best AI video generator with sound 2026 covers the wider field of models that generate audio natively.
Native audio changes how you draft, not just what you deliver. Run Gemini Omni Flash from 51 credits on the 8frame canvas alongside every other major model, from $19/month.