Google's Omni 1.1 Flash is getting 1080p and 4K video output

Google's DeepMind page for Gemini Omni now lists the model id gemini-omni-1.1-flash alongside a resolution ladder: draft at 360p for speed, generate at standard 720p, then upscale to 1080p or 4K for "production-ready outputs." Telegram chatter says the 1080p/4K options are rolling out shortly — the docs are already there, the toggle isn't everywhere yet.

The ladder is the interesting bit, not the pixel count. Omni Flash was built as the cheap, conversational half of the Omni family — you throw text, an image, audio, or a clip at it and edit the result by talking to it. Charging draft-tier compute for the 20 iterations nobody sees, and full freight only for the one you export, is how you make that loop affordable. Public preview in June shipped 720p only, so this closes the gap between "fun to prompt" and "usable in a deliverable."

Worth keeping expectations calibrated: 4K here is an upscale, not native generation, and reviewers testing Omni Flash found its textures still short of cinema grade. You get delivery-ready dimensions, not automatically delivery-ready footage. Clip lengths also remain short — the first release capped out around 10 seconds.

Why it matters: resolution was the last excuse to keep AI video in the mood-board phase instead of the timeline.