Guide2026/08/28

Draft at 360p, Upscale to 4K with Gemini Omni 1.1 Flash

Draft Gemini Omni 1.1 Flash at 360p ($0.03/s), then upscale to 1080p or 4K. Cost tables, when upscaling is worth it, and the 720p default.

Draft at 360p, Upscale to 4K with Gemini Omni 1.1 Flash

Quick Answer

Gemini Omni 1.1 Flash is cheapest at 360p ($0.03 per second), default at 720p ($0.10), and finishes at 1080p ($0.15) or 4K ($0.30) via upscaling — not native 4K generation. Google's pitch is the draft loop: iterate up to 60% faster at one-third the 720p cost, then upscale the keeper. A 10-second idea you re-roll three times at 720p is $4.00; the same three drafts at 360p plus one 720p keeper is $1.90. This page is the cost-and-workflow companion to the 1.1 announcement — the ladder, the loop, when 4K is worth $0.30/s, and when to stop at 720p.

Key Takeaways

  • Four rungs, one model. 360p / 720p / 1080p / 4K on gemini-omni-1.1-flash. Preview-era Omni Flash was 720p only.
  • 1080p and 4K are upscales. Native generation stays 720p (or 360p drafts). You are not getting a native 4K diffusion pass.
  • 360p is the iteration lane. Up to 60% faster, ~⅓ the 720p price. Use it for "does this camera move work," not for delivery.
  • Pay for output seconds, every time. Edits, extensions, and upscales that re-render seconds all bill. Polish the story at 360p; spend 4K once.
  • Shorts don't need 4K. 9:16 social compresses 4K back down. 4K is for the cut that will be graded, broadcast, or zoomed.

The 1.1 resolution ladder

From Google's published per-second rates for Gemini Omni 1.1 Flash:

ResolutionPer second5s clip10s clip40s extended take
360p (draft)$0.03$0.15$0.30$1.20
720p (default)$0.10$0.50$1.00$4.00
1080p (upscale)$0.15$0.75$1.50$6.00
4K (upscale)$0.30$1.50$3.00$12.00

Input is $1.50 per 1M tokens across text / image / video / audio. The number that moves your bill is output seconds × rung.

The leaked 1.1 pricing table we tracked before launch was this ladder — new rungs, not a 720p price cut. Headline 720p is still ~$0.10/s, same as the June preview and in the same neighborhood as Veo 3.1 Fast.

The draft → keep → upscale loop

This is the workflow Google demoed (including in Flow on a phone): cheap, then sharp.

  1. Lock the idea at 360p. Text-to-video, image-to-video, first-last interpolation, even the first scene-extension step. You are testing prompt, camera, and continuity, not delivery grain.
  2. Re-roll only the 360p misses. Three 10-second drafts = $0.90. Three 720p drafts = $3.00. The delta is the whole point of the rung.
  3. Conversational-edit on the cheap clip. "Slow the dolly." "Warm the key." Each turn bills at the current resolution. Editing a 360p 5-second take three times is $0.60, not $2.00.
  4. Re-render or upscale the keeper. Once the 360p take is the one, request 720p (default delivery) or 1080p / 4K for the finish. Don't upscale a clip you'd still re-roll.
  5. Extend after the story is right, not after 4K. A 40-second 4K chain is $12. A 40-second 360p chain is $1.20. Extend in draft, upscale the finished file.

Worked example — a 10-second hero shot, 3 drafts, 2 edit turns, one keep:

PathMathTotal
Old habit (all 720p)3 × $1.00 drafts + 2 × $1.00 edits$5.00
1.1 draft loop3 × $0.30 drafts + 2 × $0.30 edits + 1 × $1.00 keep$2.50
Draft loop + 4K finish$2.50 − $1.00 keep + $3.00 4K$4.50

You can spend more than the old path if you 4K every experiment. The savings only show up if 360p is where the failures happen.

When each rung is the right one

360p — prompt tests, storyboard animatics, first-last camera-path checks, the first two extension steps. Anything you might throw away.

720p — default delivery. Most YouTube and web embeds, most social (before the platform recompresses anyway). Stop here unless a client or a grade demands more.

1080p — the "looks sharp on a laptop" finish. ~1.5× 720p. Sensible for a hero you will cut in Premiere without a 4K timeline.

4K — delivery masters, punch-ins, broadcast, client specs that say 4K. You are paying 3× 720p for an upscale. If the 720p already has texture crawl or identity flicker, 4K will sharpen the artifact. Fix the take first.

9:16 Shorts, Reels, TikTok: stay at 720p unless you need a master for a larger cut. The apps will squash 4K. Save the $0.20/s.

How to set resolution in the API

Model ID gemini-omni-1.1-flash. The resolution lives on the response format (Google's announcement sample uses this for a 360p extend):

from google import genai

client = genai.Client()

interaction = client.interactions.create(
    model="gemini-omni-1.1-flash",
    input="A marble rolling the length of a timber track, continuous smooth shot.",
    response_format={
        "resolution": "360p",  # 360p | 720p | 1080p | 4k
    },
)

Omit response_format and you get 720p. 4k takes longer; budget wall-clock as well as money. Setup and the rest of the call shape: API tutorial.

What upscaling does not fix

  • A bad prompt. 4K of a soggy camera move is a sharper soggy camera move.
  • Identity drift. If the face already slipped at 360p, upscale will not put the right person back. Re-generate with a character reference.
  • The 10-second cap. Resolution is orthogonal to length. 4K is still 3–10 seconds per generation; extend for 40, chain past that.
  • Native detail that was never there. Google is clear: 1080p and 4K are upscaled from the model's native pass. Treat 4K as a finishing filter, not as a different model.

FAQ

How much does Gemini Omni 1.1 Flash cost per second?

$0.03 at 360p, $0.10 at 720p, $0.15 at 1080p, $0.30 at 4K. 720p is unchanged from the 1.0 preview. The new story is the cheap draft rung and the paid upscale rungs.

Is 4K native on Omni 1.1 Flash?

No. Default generation is 720p (or 360p if you ask). 1080p and 4K are upscaled outputs. Plan quality at 720p; use 4K when the take is already right.

Should I generate everything at 360p?

Generate experiments at 360p. Deliver at 720p or above. A 360p file in a 4K timeline will look like a 360p file.

How much is a 40-second 4K Omni video?

$12.00 raw at $0.30/s, before retakes. Draft that same 40 seconds at 360p for $1.20, then upscale the keeper.

Does Omni Flash the product use these Google prices?

Omni Flash is an independent platform and sells credits, not Google's meter. The table above is Google's API list price for anyone calling gemini-omni-1.1-flash directly. Credit packs on this site are on the pricing page.

Is Omni Flash affiliated with Google?

No. Omni Flash is an independent platform built for the Gemini Omni era. It is not affiliated with Google. Gemini, Gemini Omni, Omni Flash, Veo, and related names are trademarks of their respective owners.

Next steps