Skip to main content
SunoMV SunoMV

Gemini Omni 1.2 Flash: Longer Shots, Fewer Jump Cuts, Lower Price

Google released Gemini Omni 1.2 Flash on 5 November 2026. In SunoMV the Gemini Omni row now runs 1.2: text or image in, first-and-last-frame transitions between scenes, a reference image for your lead, 3–15 second clips from a 360p draft up to 4K. We cut the row's price by 20% the same day.

3–15s shots 360p → 4K 20% cheaper
Add SunoMV as a preferred source on Google See more SunoMV in Top Stories and AI answers.
Model
Gemini Omni 1.2 Flash
Shot length
3–15 s (frame pair 3–10 s)
Resolution
360p draft · 720P · 1080P · 4K
Measured run
15.04 s clip in 58 s (360p)

Use it when one take has to carry a whole lyric line.

Omni 1.2 keeps everything 1.1 did in SunoMV — text, image and first-and-last-frame shots — and adds the 15-second take, a lower price and better scene-to-scene continuity. Pin a reference image of your lead, draft at 360p, and re-run the shots that hold at 720P or above. If you are still hunting a look for the whole song, board it on a cheaper row first.

Made in SunoMV · real exports

See what a song turns into

Every sample below is an untouched SunoMV export: 3 AI storyboard shots, one AI-filmed clip, then the lyrics. Each one uses a different caption style and frame size.

All 13 samples →
Make one like this Free to start. No editing skills needed.

Features

What changed from Omni 1.1

1.2 is not a new product. It is the same Omni row with three things a music video actually feels: length, price and continuity.

Up to 15 seconds per clip

Generation runs 3–15 seconds, 10 by default. A full chorus line can sit in one take instead of two stitched clips. First-and-last-frame transitions are the exception: Google caps those at 3–10 seconds.

About 30% cheaper at the API

Google's list price at 720p is about $0.07 a second, down from about $0.10 on 1.1 — roughly $0.70 for a 10-second clip. SunoMV passed part of that on: the Omni row costs 20% less than before.

Fewer jump cuts between scenes

Google's notes for 1.2: better temporal coherence across scenes, fewer jump cuts and visual artifacts, improved character voices, and closer adherence to the reference images you upload.

What the SunoMV row does today

Google's model page and the SunoMV picker are not the same list. This is the picker.

Text, image, or two frames

Write the shot, start from a still, or pin a first and a last frame so one MV scene walks into the next. Add a reference image of your lead so the same person shows up shot after shot.

Draft, 720P, 1080P or 4K

The draft tier is labelled 480P in the picker and renders at 360p — the cheap way to check a cut. 720P is native. Google lists 1080p and 4K as upscaled outputs. Aspect ratio is 16:9 or 9:16.

Not in this picker: editing and extension

Google's API also edits clips up to 15 seconds and extends a clip 3–10 seconds per turn, up to 30 seconds total. The SunoMV row does not expose those two. Generate separate shots and cut them on the lyric timeline.

Make the shots inside SunoMV

No developer key, no waiting list. The model is already in the video picker of the music-video flow.

Pick Gemini Omni in the list

Open the create flow and choose Gemini Omni. Set 3–15 seconds, a resolution tier, and 16:9 or 9:16. That is the whole setup.

Draft at 360p, keep at 720P or above

Our 360p test: a 15-second text-to-video request came back as a 15.04-second clip with audio in 58 seconds. Check the cut cheap, then re-run the keepers at a higher tier.

Stack shots into the video

Drop each clip on the lyric timeline, preview vertical or widescreen, and export when the chorus cuts clean. Previewing the assembly is free.

From a lyric line to a stacked preview, three steps

Nothing to install. Gemini Omni is already in the SunoMV video picker.

  1. 1

    Pick Gemini Omni

    Open create, choose Gemini Omni, set 3–15 seconds, a resolution tier (360p draft, 720P, 1080P or 4K) and 16:9 or 9:16.

  2. 2

    Prompt the shot, or pin both frames

    Write the scene, start from a still, or set a first and last frame (3–10 seconds) so two scenes connect. Add a reference image of your lead to keep the same face.

  3. 3

    Stack, preview, post

    Lay the clips on the lyric timeline. Preview free, then export vertical or widescreen for TikTok, Reels, Shorts and YouTube.

When Omni 1.2 is the right row

No video model is better at everything. These are the jobs this row is for.

A chorus line in one take

Fifteen seconds covers most chorus lines. One take means one camera and one face, instead of two clips that have to be matched by hand.

Scene-to-scene transitions

Pin the last frame of scene A and the first frame of scene B. The model fills the 3–10 seconds between them, so the cut reads as a camera move instead of a jump.

A 4K finish for the keepers

Check the cut at the 360p draft tier, then render the two or three shots that matter at 1080P or 4K for a full-screen phone or a TV.

Omni 1.2 next to the rows around it

Same picker, different jobs. Click through instead of guessing.

Gemini Omni 1.2 Flash, answered

What is Gemini Omni 1.2 Flash?

Gemini Omni 1.2 Flash is a Google video model released on 5 November 2026 that generates and edits clips of up to 15 seconds from text, images or a first-and-last-frame pair. It costs about 30% less than version 1.1 through the API and is also available in Flow, Google Vids and the Gemini app.

What is first-and-last-frame interpolation?

First-and-last-frame interpolation is a way of generating video where the user supplies the opening picture and the closing picture and the model creates the motion in between. In a music video it joins two scenes with a camera move, so the same character and setting carry across the cut. Omni 1.2 allows 3–10 seconds for this mode.

What does upscaled 1080p or 4K mean?

Upscaled output means the model renders the clip at a lower native resolution and then enlarges it to 1080p or 4K. For Gemini Omni 1.2 Flash, Google prices 720p as the native tier and lists 1080p and 4K as upscaled, at about $0.104 and $0.209 per second of video.

Sources

Google figures are from the 5 November 2026 release and the AI Studio docs. SunoMV figures describe the picker after the same-day upgrade. Timings are our own runs, not vendor numbers.

  • Google released Gemini Omni 1.2 Flash on 5 November 2026: video generation and editing up to 15 seconds, better coherence across scenes, fewer jump cuts, about 30% lower API price than 1.1. Available in Flow, Google Vids, the Gemini app and the API.

    Google AI on X — Gemini Omni 1.2 Flash announcement ↗
  • Model id gemini-omni-1.2-flash. Duration 3–15 seconds (default 10); first-and-last-frame 3–10 seconds; extension 3–10 seconds per turn up to 30 seconds; up to 6 reference images (3 per character, 2 characters); 16:9 and 9:16; 1.1 requests route to 1.2 from 3 December 2026.

    Google AI Studio — Omni docs ↗
  • About 30% below 1.1 (720p was about $0.10 a second). Per second: 360p $0.023, 720p about $0.07, 1080p upscaled $0.104, 4K upscaled $0.209 — about $0.70 for a 10-second and $1.04 for a 15-second 720p clip.

    Google AI Studio — Pricing ↗
  • Our runs on 5 November 2026 at the 360p draft tier: first-and-last-frame, 5 s requested → 5.03 s clip with audio in 37 s; text-to-video, 15 s requested → 15.04 s clip with audio in 58 s. The SunoMV Omni row was upgraded to 1.2 and its price cut by 20% the same day.

    SunoMV — measured runs 2026-11-05 ↗

Frequently Asked Questions

Ask us anything!

Related guides

Pick Gemini Omni and cut the next lyric shot

Up to 15 seconds a take, first and last frame where scenes connect, 20% cheaper than before. Preview the full-song assembly free, then export for TikTok, Reels, Shorts and YouTube.