Gemini Omni 1.2 Flash’s Real Discount Is the 360p Draft, Not the 30% Price Cut

A computer monitor displaying video editing software with a timeline and preview window

Google’s own pricing page carries the footnote that complicates its headline. Gemini Omni 1.2 Flash, released on November 5, renders 720p video at “approximately $0.07 per second,” which works out to “~$1.04 for a 15s clip.” Two rows down, the older Omni 1.1 is listed at about $0.10 per second, so a full 10-second clip on the old model cost roughly $1.00. The 30% price cut is real, and so is the fact that a maxed-out clip now costs about the same as a maxed-out clip did in August. You just get five more seconds for it. (Gemini API pricing)

That is not a complaint. Fifteen seconds is the difference between a shot and a scene, and partner invideo told Google Cloud that the new length fits a full scene in one generation (eesel). But if you came for the discount, the bigger saving in this release isn’t the per-second rate. It’s the 360p draft tier, and what Google now lets you do between the draft and the final.

What changed from Omni 1.1

Google’s launch post, written by DeepMind product manager Anish Nangia, lists the headline changes:

  • Clip length: 3 to 15 seconds, set in one-second steps (1.1 stopped at 10).
  • Price: video output drops from $17.50 to $12.00 per million tokens, which is where the 30% comes from.
  • Multi-turn editing: swap a product or background, change lighting or weather, or switch the aspect ratio “turn by turn,” across the full 15 seconds.
  • References: combine character, product or style images with motion clips up to three seconds long and audio cues.
  • Audio: cleaner lip-sync and separate conditioning for multiple speakers.
  • Watermarking: every output carries an invisible SynthID mark.

Google also lists what still breaks: “fast sports action, rapid script changes and complex physics.” One of its own side-by-side demos shows a soup ladle turning into chopsticks. If your channel lives on action shots, this is not the model that fixes that.

There is one quiet step backward. The model docs cap per-turn extension at 3 to 10 seconds and 30 seconds total. eesel notes that 1.1 advertised 40 seconds of total extension. So if you were stitching long takes out of extensions, budget for a cut.

The draft tier is the real discount

Video is billed per output token, and resolution decides how many tokens a second costs. Here is Google’s footnote turned into per-clip prices:

Resolution Per second 15-second clip Notes
360p $0.023 about $0.35 Draft quality
720p about $0.07 about $1.04 Native output
1080p $0.104 about $1.56 “With upsampling”
4K $0.209 about $3.13 “With upsampling”

Per-second rates are from the Gemini API pricing page; the 15-second figures are simple multiplication, and braindetox reached the same numbers.

The workflow Google describes is “draft fast at 360p or 720p, fix a detail with a follow-up edit instead of re-generating the entire clip, and then upscale your best creations to 1080p or 4K.” Run the numbers on a typical session where you need five attempts to land a shot:

  • Five takes at 720p: about $5.20.
  • Five drafts at 360p, then one 720p render of the winner: about $1.75 plus $1.04, so roughly $2.80.
  • Four drafts plus one follow-up edit instead of a fifth re-roll: cheaper again, since the edit keeps the parts you liked.

That is close to a 50% saving from changing your habits, on top of the 30% Google already took off. It’s the same lesson we saw with Krea 3’s fast edits: when iteration gets cheap, you stop cramming five changes into one expensive prompt.

Two caveats before you rebuild your pipeline around it. First, the docs don’t say whether upscaling a draft is billed on top of the draft you already paid for, or whether the 1080p price simply replaces it; eesel flagged the same gap. Second, a draft is a preview, not a lock. eesel quotes a Hacker News report that drafts and final renders don’t come out identical, so the 720p version of your favorite 360p take may still drift. Treat the draft as a cheap way to reject bad ideas, not a guarantee of the final frame.

There is also no free tier on the API at all, for either version. And if you have scripts calling gemini-omni-1.1-flash, they get rerouted to 1.2 on December 3 and billed at 1.2 rates, so check your clip lengths before your bill does.

Where you can use it without paying per second

Most solo creators won’t touch the API. The more interesting line in the launch is that Google Vids generates 15-second clips “at no additional cost” with a Google account. The Workspace Updates post confirms the consumer list includes “users with personal Google accounts” alongside Google AI Pro and Ultra, plus most Business, Enterprise and Education plans. Rollout started November 5.

What it doesn’t publish is how many clips you get. The post points to a Vids AI limits help page and says only that the AI Expanded Access add-on gets “higher limits.” Assume a free account means a small daily allowance, not a production budget.

The other surfaces split by price:

  • Google Vids: 15-second clips, a 3 to 15 second slider, no extra cost on eligible accounts. Best place to test it.
  • Flow: directing scenes and multi-turn edits; this is where the editing loop lives for non-developers.
  • Gemini app: 15-second videos rolling out to Google AI Ultra subscribers only.
  • Gemini API and Vertex AI: pay per token, as above.

If you tried the first Gemini Omni Flash launch back in May, 1.2 is worth a fresh look mainly for the length and the edit loop. If you’re choosing between models for longer clips, our Kling 4.0, Wan 3.0 and Seedance 2.5 comparison covers the 30-second options, where a failed take costs twice as much time. And as with OpenAI’s Images 2.6, the per-unit price is only half the story; what you actually pay depends on how many units each attempt burns.

The practical move this week: open Vids on a personal account, generate a 15-second clip, then try the same prompt in Flow with two follow-up edits. You’ll know within an hour whether the edit loop saves you re-rolls, and that matters more than the 30%.

Ty Sutherland

Ty Sutherland is the Chief Editor of Full-stack Creators. Ty is lifelong creator who's journey began with recording music at the tender age of 12 and crafting video content during his high school years. This passion for storytelling led him to the University of Regina's film faculty, where he honed his craft. Post-university, Ty transitioned into the technology realm, amassing 25 years of experience in coding and systems administration. His tenure at Electronic Arts provided a deep dive into the entertainment and game development sectors. As the GM of a data center and later the COO of WTFast, Ty's focus sharpened on product strategy, intertwining it with marketing and community-building, particularly within the gaming community. Outside of his professional pursuits, Ty remains an enthusiastic content creator. He's deeply intrigued by AI's potential in augmenting individual skill sets, enabling them to unleash their innate talents. At Full-stack Creators, Ty's mission is clear: to impart the wealth of knowledge he's gathered over the years, assisting creators across all mediums and genres in their artistic endeavors.

Recent Posts