Skip to content
audiotale.net
Sign in
NewsNews

Gemini Omni 1.2 Flash: 15-second AI video at $0.07 per second, 30% cheaper

Google made its Omni video model generally available with 15-second clips, multi-turn editing and a 30 percent price cut, but 1080p and 4K prices are not in the launch posts.

audiotale.net team

· 6 min read

Written by AI from the sources linked below. How we use AI

On this page
  1. Key points
  2. What happened
  3. Background
  4. The numbers
  5. What it means
  6. What we don't know yet
  7. What to watch next
  8. Sources

Google made Gemini Omni 1.2 Flash, its AI video model, generally available on 5 November 2026. Clips can now run up to 15 seconds (the previous version stopped at 10), a generated clip can be edited over several chat turns, and the price of 720p output fell 30 percent to $0.07 per second. For developers and teams that produce product videos, ads and try-on clips in volume, the change is mostly about cost per finished clip and about not having to regenerate a whole video to fix one detail.

Key points

  • Gemini Omni 1.2 Flash became generally available on 5 November 2026 through the Gemini API and, per Google Cloud's post of 6 November, on Google Cloud for the first time for Omni.
  • Clips can be 3 to 15 seconds long in one-second steps, up from a 10-second maximum on Omni 1.1 Flash.
  • Google lists 720p video at $0.07 per second, 30 percent less than Omni 1.1 Flash, and says other resolutions get "similar savings".
  • Clips can be refined turn by turn (swap a product or background, change lighting or aspect ratio) across the full 15 seconds, and every output carries an invisible SynthID watermark.
  • Google says it is still working on fast sports action, rapid script changes and complex physics.

What happened

Google DeepMind announced Gemini Omni 1.2 Flash on 5 November in a post by product manager Anish Nangia. It describes four headline changes: native 15-second video, better adherence to reference material, multi-turn conversational editing and a lower price. The model is generally available on the Gemini API and on Google Cloud, which Google's cloud team says is the first time Gemini Omni has been generally available there, backed by Provisioned Throughput (reserved capacity) for production workloads.

The length change is the easiest to state. Duration can be set in one-second increments from 3 to 15 seconds. Google's examples include a four-shot, 15-second science explainer generated from a single style frame, and a continuous drone-style flythrough with no cuts.

Reference handling is the second change. A request can combine character, product or style images, short motion clips of up to three seconds, and audio cues, and Google says faces, garments, packaging typography and brand colours stay consistent across camera angles. Google's Cloud post adds a caveat that audio references are gated behind allowlisting, so not every account can use them yet.

The third is editing. Instead of re-rolling a clip when one thing is wrong, you continue the same conversation: swap a product, change the weather or the lighting, switch the aspect ratio. Google says this works across full 15-second clips and shows an example where the background changes through five scenes while the characters' performances and camera framing stay intact.

Google also describes a cheaper workflow built around the model: draft at 360p or 720p, fix details with follow-up edits, and upscale only the final version to 1080p or 4K.

Availability outside the API is staggered:

  • Google Vids: Google's Workspace Updates blog says Omni 1.2 is rolling out to Vids from 5 November, with full rollout taking 1 to 3 days for Rapid Release and Scheduled Release domains. There is no admin control for it. Google DeepMind's post says 15-second clips in Vids come at no additional cost with a Google account or Workspace subscription.
  • Flow, Google Ads and Gemini for business: listed as rolling out.
  • Gemini app: 15-second videos are rolling out to Google AI Ultra subscribers.

Background

Google first showed Omni Flash at its I/O conference in May 2026, with clips capped at 10 seconds, according to The Next Web. It reached developers in June at $0.10 per second of 720p video, per the same report, and Omni 1.1 Flash kept that ceiling of 10 seconds. The new price follows directly: $0.10 less 30 percent is $0.07, which matches Google's figure.

The competitive picture matters too. The Next Web notes that ByteDance's Seedance 2.5, unveiled in June, makes 30-second clips, so Google is not the leader on length. Its pitch is instead cost, consistency and integration with products people already use, from Vids to Ads. We have not independently tested any of these models, and the comparison below is about published limits only.

The same announcement that touts "less visual drift or morphing" also lists what is not solved: fast sports action, rapid script changes and complex physics. Anyone building a pipeline should treat those as the areas to test before committing.

The numbers

The price cut is easy to see in a per-clip calculation. The figures below are our arithmetic from the published 720p list prices, not figures Google published for whole clips:

  • A 10-second clip on Omni 1.1 Flash at $0.10 per second: $1.00.
  • A 10-second clip on Omni 1.2 Flash at $0.07 per second: $0.70.
  • A 15-second clip on Omni 1.2 Flash: $1.05.

So the longest new clip costs about the same as the longest old clip, while a clip of the old maximum length costs 30 percent less. The Google Cloud post says other resolutions see "similar savings", but it does not list the 1080p or 4K prices in the text we read, so we do not give them here. Check the Gemini API pricing page for current numbers before budgeting.

Gemini Omni Flash price per second of 720p videoUS$ per second
Gemini Omni Flash price per second of 720p video (US$ per second)
US$ per second
Omni 1.1 Flash0.10
Omni 1.2 Flash0.07

Source: Google DeepMind (1.2 Flash price and 30% cut); The Next Web (earlier $0.10)

Maximum clip length per generationseconds
Maximum clip length per generation (seconds)
seconds
Seedance 2.530
Omni 1.2 Flash15
Omni 1.1 Flash10

Source: Google DeepMind (Omni 1.1 and 1.2 Flash); The Next Web (Seedance 2.5)

What it means

For developers. The 15-second ceiling removes a common workaround: stitching two 10-second clips and fighting continuity at the seam. Multi-turn editing changes the cost model, because a one-line correction should be cheaper than a full regeneration, though Google does not publish the price of an edit turn in the material we read. Provisioned Throughput and general availability on Google Cloud matter for teams that want predictable capacity rather than preview limits.

For marketing and e-commerce teams. Google's examples are all catalogue and ad work: virtual try-on sequences from a model photo plus product shots, product spins, ad variants for different placements. Photoroom, a customer quoted in Google Cloud's post, said version 1.2 brings 15-second videos, "a top user request". Companies make such statements at launch, and they are not an evaluation, but they show where the model is being aimed.

For creators and Workspace users. In Vids the model is included in the product rather than billed per second, subject to the plan's AI limits, which Google documents separately.

For trust and safety. Every output carries an imperceptible SynthID watermark, according to Google. That helps platforms and tools that can detect it, but it is not visible to viewers, and it does not stop someone passing a clip through a tool that strips or degrades it. Treat it as one signal, not a guarantee.

The Next Web reports that, in Google DeepMind's own comparisons, testers preferred 1.2 over 1.1 for visual consistency 62 percent of the time. We could not find that figure in the Google posts we read, so we attribute it to The Next Web and note it is a vendor-run comparison.

What we don't know yet

  • Prices for 1080p and 4K, and the cost of an edit turn, are not stated in the sources we read. Google says savings at other resolutions are similar.
  • How the model performs on independent video leaderboards. No third-party evaluation of 1.2 Flash was available when we wrote this.
  • How many Gemini app users beyond Google AI Ultra will get 15-second clips, and when.
  • Which accounts have audio references enabled, since they require allowlisting.
  • Whether the 62 percent consistency preference holds up outside Google's own tests.

What to watch next

Watch for independent benchmark results and community comparisons against Seedance and other video models, for the Gemini API pricing page to show the full resolution ladder, and for whether Google widens the Gemini app rollout beyond the Ultra tier. Also watch how SynthID detection is exposed to third parties, since more 15-second ads and product clips will be circulating soon.

Sources

SaveStart the discussion

Written by

audiotale.net team

The official account of the people who run audiotale.net: announcements, conversation starters and the newsroom's drafts. Posts written with AI say so. Every other account here is a real person.

Read next