How Much Does Gemini Omni 1.2 Flash Cost Per Clip?

How much does Gemini Omni 1.2 Flash cost per clip?

Gemini Omni 1.2 Flash costs about 7 cents per second at 720p. Google says that is 30% cheaper than the previous version. Clips can run anywhere from 3 to 15 seconds, so a clip at 720p costs roughly $0.21 at the short end and $1.05 for a full 15 seconds. You can draft at low resolution and then upscale to 1080p or 4K. The sources I have do not list per-second prices for those higher resolutions.

What is Gemini Omni 1.2 Flash?

Gemini Omni 1.2 Flash is Google's updated video generation model. It is generally available, and Google positions it as a large step up from the prior version. The pitch centers on stronger consistency across shots, more natural multi-speaker dialogue, smoother camera moves, and editing through conversation.

The clips come with sound. Demonstrations from Google included a 15 second animated scene where four characters each spoke in a matched voice with lip sync, and a product explainer with on-screen call-outs that stayed readable while things moved.

What does a clip cost at each length?

The only per-second price in the source is the 720p rate of 7 cents. Multiplying it out gives these figures. They are straight arithmetic, not a published rate card.

Because you can choose any exact length between 3 and 15 seconds, you pay only for what you ask for. A 7 second shot costs $0.49. You do not have to round up to a fixed bucket.

The 30% discount also lets you back into the old price. If 7 cents is 30% below the previous rate, the earlier price was about 10 cents per second. Treat that as an estimate, since the figure comes from the math and not from a stated number.

How long can Omni 1.2 Flash clips be?

Clips can now run up to 15 seconds, up from 10 in the previous version. The minimum is 3 seconds. Within that range you set the length yourself.

The extra five seconds matter for single-take shots. One test prompt asked for a continuous 15 second drone move from the summit of K2 down a glacier valley and into an alpine meadow, with no cuts. Another Google example described one FPV path across a piazza, through bronze doors and up through a dome. A longer maximum gives a continuous move more room to play out before you need to stitch clips together.

What resolution options does it offer?

The base price applies to 720p. From there, you can upscale to 1080p or 4K. The sources do not break out what each upscale step costs, so I will not guess at it.

Resolution is tied to the workflow below, which is where most of the real cost control happens.

How does the draft-then-upscale workflow work?

The idea is simple. Generate at low resolution, judge the result, and only spend more on the clip you want to keep.

1. Write the prompt and generate at 720p. At 7 cents per second, a 15 second draft costs about $1.05.
2. Check the output for motion, consistency, dialogue and framing. Iterate on the prompt or edit through conversation if something is off.
3. Once a clip is right, upscale that single clip to 1080p or 4K.

The savings come from not paying high-resolution rates for clips you throw away. Video generation is trial and error. Five 15 second drafts at 720p come to roughly $5.25 by the same arithmetic. Running those five attempts at a higher resolution would cost more, though the exact multiple depends on pricing the sources do not list.

Multi-turn editing fits the same pattern. In one demonstration, a clip's background was swapped over successive prompts, from a pavilion tent to an oak forest stable with a white horse. Each step is a chance to refine before you commit to a final render. Whether each edit is billed as a fresh generation was not covered, so confirm that before you plan an iterative workflow.

How do you access Gemini Omni 1.2 Flash?

There are three routes mentioned:

• Google AI Studio. Select the Gemini Omni model at the top and enter your prompt. This suits quick testing.
• The API. Install Google's SDK, get an API key from AI Studio, and send a prompt from a short script. This is the route to use for serious or repeated generation.
• Vertex AI on Google Cloud. Some of Google's own showcase examples were run here.

For builders, the API is the practical choice. Per-second billing is easy to track in code, and you can set the duration per request.

Is Omni 1.2 Flash worth the price?

It depends on what you compare it to. A $1.05 full-length clip is cheap for something that arrives with synced audio, multiple speakers and a continuous camera move. It is not free, though, and costs add up when you iterate.

Where it makes sense:

• Short social or ad clips where 3 to 15 seconds is the target length.
• Prototyping storyboards, since you can draft cheaply at 720p.
• Dialogue scenes where voice-to-character matching would otherwise need separate tools.

Where to be careful:

• High-volume 4K output, because upscale pricing is not confirmed in the sources.
• Workflows that need many retries, since drafts are billed too.
• Anything needing pixel-exact brand or product fidelity. Google's try-on and poster examples show improved consistency, but they are curated demos, so test with your own assets.

Running local video models is the alternative. The creator who demonstrated the model argued that local options still cost a lot of time, energy and hardware to get right. A per-second API trades that effort for a metered bill.