Google quietly pushed out version 1.1 of its Gemini Omni Flash video model, and the headline feature isn’t raw quality — it’s economics. Draft-mode generation at 360p now runs up to 60 percent faster and costs a third of what 720p output does. Per-second pricing starts at $0.03 for 360p, climbing to $0.10 for 720p, $0.15 for 1080p, and $0.30 for 4K. For developers building video features at scale, that’s the difference between a proof-of-concept and something you can actually ship.
Scene extension got a meaningful upgrade too. The model now analyzes up to ten seconds of existing footage when extending a scene, rather than just the final second as before. That’s a big deal for visual consistency — characters, lighting, and motion patterns have far more context to anchor to. Extensions can run in 10-second increments up to 40 seconds total. Google also added style-reference support: developers can upload up to three seconds of external footage to carry over characters or motion patterns into a new generation, plus set start and end frames to define camera movement between keyframes.
The practical workflow looks like this: generate a cheap 360p draft to iterate fast, lock in the composition, then upscale to 1080p or 4K for final delivery. That mirrors what image generators normalized years ago — rough drafts first, expensive renders only when you’re confident. Video generation has been held back by exactly this missing rung on the cost ladder. Google just built it.
Omni 1.1 is available now through Google AI Studio and the developer docs. The per-second pricing model is also a departure from the per-generation credits some competitors use, which makes cost forecasting more predictable for production workloads. Whether the 360p draft mode produces artifacts that survive upscaling gracefully remains an open question — Google hasn’t published side-by-side comparisons yet. But the direction is clear: AI video is shifting from spectacle to infrastructure.
