Found Motive

Compare

Veo 3.1 vs Gemini Omni Flash: which Google video model should you use?

Use Gemini Omni Flash for new work. Google calls it the default video model in the Gemini API and says it replaces Veo in the Gemini app. Keep Veo 3.1 only for scene extension, last-frame control or older pipelines, through Google Cloud or Flow, because Google shut down the Veo 3.1 preview models in the Gemini API on Oct 22, 2026.

By · Published · Latest claim verified . We have not tested these tools ourselves. How to read the labels.

Status checked : Google shut down the Veo 3.1 preview models in the Gemini API on Oct 22, 2026. Veo 3.1 is still in Google Flow and, as GA models, on Google Cloud.

Disclosure. Found Motive has no affiliate links on this page and earns nothing from the tools on it. Where we do have an affiliate link, it is marked next to the link and does not change what we say. How we handle affiliate links.

Where can you still use Veo 3.1 and Omni Flash?

Where each model runs, checked on Google's pages on Oct 21, 2026; Gemini API row re-checked Oct 23, 2026.
FeatureVeo 3.1Lite, Fast and QualityGemini Omni Flashgemini-omni-1.1-flash
Gemini app

Being replaced. Google says Gemini Omni will replace Veo in the Gemini app.

OFFICIAL Gemini app: video generation (opens in new tab) Verified

Omni 1.1 Flash makes 10-second videos with native audio, from text or up to 5 photos, with scene extension, video-to-video and multi-turn editing. Google AI subscription required; 18+.

OFFICIAL Gemini app: video generation (opens in new tab) Verified

Google Flow

Still listed: Veo 3.1 Lite, Fast and Quality make 4, 6 or 8-second clips.

OFFICIAL Google Flow Help: video models (opens in new tab) Verified

Omni Flash 1.1 makes 4 to 10-second clips; Google says all Flow users can access Omni models.

OFFICIAL Google Flow Help: video models (opens in new tab) Verified

Gemini API

Shut down: Google turned off the Veo 3.1 preview models (standard, Fast and Lite) in the Gemini API on Oct 22, 2026, and points to gemini-omni-1.1-flash or the GA models on the Gemini Enterprise Agent Platform.

OFFICIAL Gemini API: video generation (opens in new tab) Page dated Retrieved

gemini-omni-1.1-flash, released Aug 27, 2026, with no shutdown date listed. Google calls Omni Flash the default video model.

OFFICIAL Gemini API: deprecations (opens in new tab) Page dated Retrieved

Google Cloud

Generally available: veo-3.1-generate-001, released Nov 17, 2025, with retirement listed as "November 17, 2026 or later".

OFFICIAL Google Cloud: Veo 3.1 (veo-3.1-generate-001) (opens in new tab) Verified

Listed in preview on Google Cloud's video-generation docs. Check Google Cloud's docs before you plan production use.

OFFICIAL Google Cloud: generate videos from references (opens in new tab) Verified

Should you use Veo 3.1 or Gemini Omni Flash in 2026?

Use Gemini Omni Flash for new work. Google calls it the default video model in the Gemini API, and says Omni replaces Veo in the Gemini app. Keep Veo 3.1 for scene extension, last-frame control or legacy pipelines, and run it through Flow or Google Cloud, where the GA model stays (checked Oct 21, 2026).

OFFICIAL Gemini API: video generation overview (opens in new tab) · Gemini app: video generation (opens in new tab) · Verified

What is Gemini Omni Flash?

Gemini Omni Flash is Google's newer video model. It went GA in the Gemini API as gemini-omni-1.1-flash on Aug 27, 2026. It takes text, images and video, makes 3 to 10-second clips with audio, and can edit, extend or upscale video. Google's Gemini app release notes list its app rollout under May 19, 2026.

OFFICIAL Gemini API changelog (opens in new tab) · Gemini API: Gemini Omni Flash model page (opens in new tab) · Gemini app release notes (opens in new tab) · Verified

Is Veo 3.1 being shut down?

Only in the Gemini API, and that has happened: Google shut down the three Veo 3.1 preview models there (standard, Fast and Lite) on Oct 22, 2026, along with gemini-omni-flash-preview, and points to gemini-omni-1.1-flash. The GA Veo 3.1 model stays on Google Cloud, and Flow still lists Veo 3.1 (checked Oct 23, 2026).

OFFICIAL Gemini API: deprecations (opens in new tab) · Google Cloud: Veo 3.1 (veo-3.1-generate-001) (opens in new tab) · Google Flow Help: video models (opens in new tab) · Verified

Which is better for editing video?

Omni Flash. In the Gemini API you edit by conversation: each turn builds on the previous result through a previous_interaction_id, and you can also edit uploaded clips of up to 10 seconds. Uploaded-video edits aren't available in the EEA, Switzerland or the UK, and voice editing isn't supported. Flow lists no video-to-video for Veo 3.1.

OFFICIAL Gemini API: Gemini Omni Flash guide (opens in new tab) · Google Flow Help: video models (opens in new tab) · Verified

Which is better for extending a clip?

Veo 3.1, for now, but no longer through the Gemini API, where its preview shut down on Oct 22, 2026. The GA Veo 3.1 model on Google Cloud lists video extension, and in Flow Veo 3.1 Lite extends clips. Omni makes a 3 to 10-second continuation from the end of a clip in the API, while Flow still marks Omni extend as coming soon (checked Oct 23, 2026).

OFFICIAL Gemini API: Veo 3.1 (opens in new tab) · Gemini API: Gemini Omni Flash guide (opens in new tab) · Google Flow Help: video models (opens in new tab) · Verified

Do Veo 3.1 and Omni Flash both make audio?

Yes. Veo 3.1 generates native audio, and Google says Omni Flash text-to-video generates a video with audio. Omni doesn't support uploaded audio references or voice editing. Flow adds preset custom voices, with some voice features limited to certain regions.

OFFICIAL Gemini API: Veo 3.1 (opens in new tab) · Gemini API: Gemini Omni Flash guide (opens in new tab) · Google Flow Help: video models (opens in new tab) · Verified

Which has higher video quality?

Both list 4K, but differently. Veo 3.1 generates 720p, 1080p or 4K. Omni Flash defaults to 720p, and Google says its 1080p and 4K outputs are upscaled. Flow also has a 360p draft mode for Omni. We haven't run side-by-side quality tests.

OFFICIAL Gemini API: Veo 3.1 (opens in new tab) · Gemini API: Gemini Omni Flash guide (opens in new tab) · Google Flow Help: video models (opens in new tab) · Verified

How do you migrate from Veo 3.1 to Omni Flash in the Gemini API?

Switch calls to gemini-omni-1.1-flash; the Veo 3.1 preview models stopped working in the Gemini API on Oct 22, 2026. Keep clips at 10 seconds or less, test edits as multi-turn conversations, and check region limits for uploaded-video edits. If you need longer extensions or last-frame control, move that work to Veo 3.1 on Google Cloud.

Migration checklist

  • Move any code that called the Veo 3.1 preview models (standard, Fast or Lite) to gemini-omni-1.1-flash; those models shut down on Oct 22, 2026.
  • Also move off gemini-omni-flash-preview, which shut down the same day.
  • Cap clips at 10 seconds (Omni makes 3 to 10-second clips).
  • Rebuild edit flows as multi-turn calls using previous_interaction_id.
  • Check region rules: uploaded-video edits and extensions aren't available in the EEA, Switzerland or the UK.
  • Drop uploaded audio references; Omni doesn't support them.
  • Keep long extensions or last-frame control on Veo 3.1 via Google Cloud.

OFFICIAL Gemini API: deprecations (opens in new tab) · Gemini API: Gemini Omni Flash guide (opens in new tab) · Google Cloud: Veo 3.1 (veo-3.1-generate-001) (opens in new tab) · Verified

Veo 3.1 vs Omni Flash: feature by feature

Veo 3.1 vs Gemini Omni Flash, from Google's docs (Oct 21, 2026; clip length and extension re-checked Oct 23, 2026). We have not run our own tests.
FeatureVeo 3.1Lite, Fast and QualityGemini Omni Flashgemini-omni-1.1-flash
Clip length

8-second videos, per Google's Veo 3.1 docs; 4, 6 or 8 seconds in Flow.

OFFICIAL Gemini API: Veo 3.1 (opens in new tab) Page dated Retrieved

3 to 10 seconds in the Gemini API; 4, 6, 8 or 10 seconds in Flow.

OFFICIAL Gemini API: Gemini Omni Flash model page (opens in new tab) Page dated Retrieved

Resolution

720p, 1080p or 4K.

OFFICIAL Gemini API: Veo 3.1 (opens in new tab) Page dated Retrieved

360p, 720p, 1080p or 4K at 24 fps. 720p is the default; Google says 1080p and 4K are upscaled.

OFFICIAL Gemini API: Gemini Omni Flash guide (opens in new tab) Page dated Retrieved

Audio

Native, generated audio.

OFFICIAL Gemini API: Veo 3.1 (opens in new tab) Page dated Retrieved

Text to video "generates a video with audio". Voice editing and audio references aren't supported.

OFFICIAL Gemini API: Gemini Omni Flash guide (opens in new tab) Page dated Retrieved

Inputs

Text, an image, first and last frame, and up to three reference images.

OFFICIAL Gemini API: Veo 3.1 (opens in new tab) Page dated Retrieved

Text, images and video (up to 10 seconds for editing and extension); up to 3 video references of up to 3 seconds each.

OFFICIAL Gemini API: Gemini Omni Flash guide (opens in new tab) Page dated Retrieved

Edit an existing video

Flow lists no video-to-video for any Veo 3.1 mode.

OFFICIAL Google Flow Help: video models (opens in new tab) Verified

Multi-turn edits that build on the previous result (previous_interaction_id), plus uploaded-video edits. Editing or extending uploaded videos isn't available in the EEA, Switzerland or the UK.

OFFICIAL Gemini API: Gemini Omni Flash guide (opens in new tab) Page dated Retrieved

Extend a clip

Google Cloud: the GA model lists video extension (the Gemini API preview, which extended up to 148 seconds, shut down Oct 22, 2026). Flow: only Veo 3.1 Lite extends.

OFFICIAL Google Cloud: Veo 3.1 (veo-3.1-generate-001) (opens in new tab) Page dated Retrieved

Gemini API: a 3 to 10-second continuation from the end of the clip. Flow: extend is marked "coming soon".

OFFICIAL Gemini API: Gemini Omni Flash guide (opens in new tab) Page dated Retrieved

Languages

Not stated on the pages we read.

UNVERIFIED Gemini API: Veo 3.1 (opens in new tab) Page dated Retrieved

Google says English is fully supported; other languages haven't been evaluated.

OFFICIAL Gemini API: Gemini Omni Flash guide (opens in new tab) Page dated Retrieved

Watermark

SynthID watermark on generated videos.

OFFICIAL Gemini API: Veo 3.1 (opens in new tab) Page dated Retrieved

SynthID watermark on all generated videos.

OFFICIAL Gemini API: Gemini Omni Flash guide (opens in new tab) Page dated Retrieved

Plans

In Flow on Google's plans; GA on Google Cloud. See Google's pages for current terms.

OFFICIAL Google Flow Help: video models (opens in new tab) Verified

Gemini app needs a Google AI subscription; Flow lists Omni for all users. See Google's pages for current terms.

OFFICIAL Gemini app: video generation (opens in new tab) Verified

Pick Veo 3.1 if… / Pick Omni Flash if…

Quick picks, based on the sourced claims above.
Pick…if you
Veo 3.1
  • need scene extension past 10 seconds or last-frame control, which Google's overview still points to Veo for.
  • run an existing Veo pipeline and can move it to the GA model on Google Cloud.
  • work in Flow and want Veo 3.1 Lite's extend today.
Gemini Omni Flash
  • are starting new work in the Gemini API or the Gemini app.
  • want to edit by conversation, or edit an uploaded clip of up to 10 seconds (outside the EEA, Switzerland and the UK).
  • want quick 360p drafts in Flow before a 720p version.

For longer pieces, a mixed workflow can work: draft and edit short shots in Omni, then extend the one you keep with Veo 3.1 on Google Cloud. That is our suggestion, not Google's.

Free test prompts: try both yourself

Run the same prompt in Flow with Veo 3.1 and with Omni Flash 1.1, at the same length. Google's Omni guide suggests describing a single continuous shot.

Single continuous shot (Omni guide style) UNTESTED TEMPLATE

In a single continuous shot, a barista pours latte art in a sunlit café, slow push-in on the cup, soft morning light, steam rising. No scene cuts. Sound: milk steaming, quiet café chatter.

Multi-turn edit (Omni) UNTESTED TEMPLATE

Turn 1: "A red bicycle leaning on a brick wall, gentle breeze, leaves falling." Turn 2: "Make it night, add a warm street lamp, keep the bicycle exactly the same."

Extend a clip (Veo 3.1) UNTESTED TEMPLATE

Continue the shot: the camera follows the cyclist down the hill, same lighting and colour, same character, smooth motion.

Dialogue with audio UNTESTED TEMPLATE

Two friends at a kitchen table. One says: "You finished the whole pizza?" The other laughs and says: "Technically, you said I could." Natural voices, room tone.

Put it to work

What we have not verified

FAQ

Is Veo 3.1 still available?

Yes, outside the Gemini API. Google Cloud lists veo-3.1-generate-001 as generally available, and Flow still lists Veo 3.1 Lite, Fast and Quality. Only the Gemini API preview models were shut down, on Oct 22, 2026 (checked Oct 23, 2026).

What replaces Veo 3.1 in the Gemini API?

gemini-omni-1.1-flash. Google names it as the replacement for the three Veo 3.1 preview models and for gemini-omni-flash-preview, all shut down on Oct 22, 2026 (checked Oct 23, 2026).

How long can Omni Flash videos be?

3 to 10 seconds per clip in the Gemini API, and 4, 6, 8 or 10 seconds in Flow. Veo 3.1 clips are 8 seconds per Google's Veo docs and 4, 6 or 8 seconds in Flow.

Does Gemini Omni Flash make audio?

Yes. Google's Omni guide says its text-to-video mode generates a video with audio included. It doesn't accept uploaded audio references, and voice editing isn't supported.

Is Gemini Omni Flash free?

In Flow, Google says all users can access Omni models. In the Gemini app it needs a Google AI subscription and you must be 18 or older (checked Oct 21, 2026). In the Gemini API it's paid tier only; Google's pricing page puts it at about $0.10 per second of 720p video (checked Oct 23, 2026).

Can I edit uploaded videos with Omni in Europe?

Not in the EEA, Switzerland or the UK. Google says editing or extending uploaded videos isn't available there yet, though you can still edit and extend videos the model generated.

More comparisons

All comparisons →

Get the weekly roundup

Playbooks, new features and tool comparisons, in one short email every Thursday.

Newsletter delivered by beehiiv. Prefer a plain page? Subscribe here. See our privacy page.