Google launches Gemini Omni 1.2 Flash with longer clips and conversational editing

Overview

Updated 46 minutes ago

Google's Gemini Omni 1.2 Flash generates video clips up to 15 seconds long, up from 10 seconds in the previous version, and edits them through follow-up prompts instead of full re-generation. The model costs $0.07 per second of 720p video, 30% less than Omni 1.1.

The change targets two costs that have gated short-form video production: the price of each generation and the time lost restarting a clip when one detail is wrong. Omni 1.2 Flash ships inside Google Vids and the Gemini app, not just a developer API.

Why it matters

If conversational editing holds up, an ad spot that took a day of re-rolls can be revised in minutes.

Questions about this story

Free account needed to ask — your question is kept and asked for you right after sign-up. Answers are public.

No questions yet — be the first to ask.

Key Indicators

15 seconds

Maximum clip length

Up from 10 seconds in Omni 1.1; adjustable in one-second steps from 3 to 15.

30%

Price cut from Omni 1.1

720p generation now costs $0.07 per second.

$0.07

Price per second of 720p video

Draft at 360p or 720p, then upscale to 1080p or 4K.

Free 15-second videos per month in Google Vids

Available to any Google account or Workspace subscriber.

Reference images per generation

Character, product, or style images plus short motion clips and audio cues.

Voices

Curated perspectives — historical figures and your fellow readers.

People Involved

Anish Nangia

Product Manager, Google DeepMind

Announced the launch of Gemini Omni 1.2 Flash

Organizations Involved

Google DeepMind

AI company

Developer of Gemini Omni 1.2 Flash

Google's AI research unit, which builds the Gemini model family.

Google

Technology company

Distributes Omni 1.2 Flash through the Gemini app, Vids, Cloud, and Ads

Parent company distributing the model across consumer and enterprise products.

Timeline

4 events Latest: Today

Tap a bar to jump to that date

November 2026
Google launches Gemini Omni 1.2 Flash
Today Product launch

Model generates 3- to 15-second clips and edits them through follow-up prompts.
Today
Google Vids adds free Omni 1.2 clips
Today Product update

Google account and Workspace users get three 15-second videos per month at no cost.
Today
Model card restricts speech editing
Today Safety

Google limits the model's ability to change people's speech while it studies safe use.
Today
Vercel adds Omni 1.2 to AI Gateway
Today Distribution

Developers can call the model with up to six reference images.
Today

Scenarios

Omni 1.2 Flash reaches all paid Gemini tiers

Likely Resolves by Mar 1, 2027

Discussed by: Google's launch post and Gemini app release notes

Google currently limits the Gemini app rollout to Google AI Ultra subscribers. If the model performs well in Vids and the API, Google is likely to extend it to Google AI Pro and Workspace business tiers, following its pattern of tiered Gemini releases.

Independent benchmark places Omni 1.2 Flash behind a rival on prompt adherence

Possible Resolves by Feb 1, 2027

Discussed by: Third-party reviewers and video-generation leaderboards

A Segmind review found roughly one instruction in four gets skipped, usually a background change or a final camera move. If an independent benchmark such as Artificial Analysis confirms that gap against a competitor, enterprise buyers may treat Omni as a drafting tool rather than a production model.

Google lifts the restriction on speech editing

Unlikely Resolves by Jun 1, 2027

Discussed by: Google DeepMind's model card

The model card says Omni 1.2 Flash can change people's speech but that Google is restricting the capability while it studies safe use. Lifting the restriction would require new safeguards, likely watermarking and voice-consent checks, and would open dubbing and localization workflows.

Historical Context

2 moments from history that rhyme with this story — and how they unfolded.

February 2024

OpenAI Sora preview (2024)

OpenAI previewed a text-to-video model that produced clips far longer than most tools at the time, setting public expectations for what AI video could do. Access was limited to a small group of testers and researchers.

Then

The preview drove a wave of competitor announcements and raised questions about training data and likeness rights.

Now

Video generation became a named frontier for the major AI labs, with length, consistency, and price as the main axes of competition.

Why this matters now

Omni 1.2 Flash competes on editing control and price per second, not raw clip length, a shift from the terms Sora set.

May 2024

Google Veo unveiling (2024)

Google introduced Veo at its I/O developer conference, a major text-to-video model, and made it available to enterprise customers through Vertex AI. The model targeted professional filmmakers and advertisers.

Then

Access stayed behind cloud contracts, limiting consumer reach.

Now

Google learned to ship video generation through both developer APIs and its own apps, a pattern Omni 1.2 Flash follows.

Why this matters now

Omni 1.2 Flash puts video generation in front of free Google Vids users, a broader distribution strategy than Veo's enterprise-first launch.