Veo 3.1 vs Veo 3.1 Fast vs Veo 3.1 Lite: which to choose

The short answer

• Veo 3.1 for the final version. The full version, with dialogue, effects and ambience created together with the picture, up to 4K and up to 3 reference images. In the independent rankings for animating an image it is the highest ranked of the three
• Veo 3.1 Fast for most of the work. According to Google it has the same options as Veo 3.1, reference images and 4K included, and it costs less. In the text to video rankings it sits almost level with Veo 3.1
• Veo 3.1 Lite for drafts and volume. Google describes it as its most cost-effective video model. Google documents neither 4K nor reference images for it, but here you can pick both, and both worked in our test

To give a sense of price, an 8 second clip at 720p with sound costs ₪5.8658.58 crVeo 3.1, 720p · 8s · with sound in Veo 3.1, ₪1.4114.06 crVeo 3.1 Fast, 720p · 8s · with sound in Fast and ₪0.707.03 crVeo 3.1 Lite, 720p · 8s · with sound in Lite right now. The prices are computed as the page loads, with the same calculation the studio shows before every generation.

The three versions side by side

The table sums up what Google states about each version in its official documentation, as we read it on October 11, 2026, and what you can pick here where that differs.

Three more things from the documentation: in all three versions an end image works only together with a start image. Google writes that a prompt in English is fully supported, while other languages have not been evaluated and may work with varying results. And as for 4K in Fast, it appears in the Gemini API documentation but not on Google’s Vertex AI model page.

One more thing we checked ourselves: with sound on, 1080p and 4K are made here by enlarging a 720p clip (in Veo 3.1, for a clip without reference images). In our test, Fast without sound made 1080p at full resolution.

What each version costs

The table is computed as the page loads, for an 8 second clip with sound, with the same calculation the studio shows before every generation. Right now Lite is the cheapest of the three at every resolution, 4K included, and the full Veo 3.1 is the most expensive.

Shorter is not always cheaper

Some models are priced by the second and some by the clip. In Fast, for example, a 4 second clip at 720p costs ₪1.4114.06 crVeo 3.1 Fast, 720p · 4s · with sound right now, and an 8 second clip costs ₪1.4114.06 crVeo 3.1 Fast, 720p · 8s · with sound. When the price is the same, there is no point in going shorter to save money: create 8 seconds and trim in editing. Every model page has a price table for every length.

Why turning sound off can cost more

In all three versions sound is on by default, and turning it off does not always save money. Right now an 8 second clip at 720p in Fast costs ₪1.4114.06 crVeo 3.1 Fast, 720p · 8s · with sound with sound and ₪3.7537.50 crVeo 3.1 Fast, 720p · 8s · no sound without it. When turning sound off costs more, the reason is that the cheapest way to create the clip always renders with sound, and a clip without sound moves to another way that costs more. The studio says so next to the switch, and if you do not need sound, you can leave it on and mute the clip.

Quality: what the independent rankings say

To compare quality, look at two independent rankings that rank video models by the votes of people who compare results: an open voting leaderboard and Artificial Analysis. This is how the rankings looked on October 11, 2026:

For text to video the differences are very small: on the open voting leaderboard Veo 3.1 has 1364 points and Fast has 1362, a gap smaller than the margin of error the leaderboard shows for each of them. For animating an image the full Veo 3.1 leads in both rankings, and on Artificial Analysis it is Lite that ranks above Fast there. Keep in mind that the rankings measure the general preference of voters, not your goal and not Hebrew. The best test is to run the same prompt in two versions and compare.

Sound and dialogue and the Hebrew question

In Veo 3.1 and Fast sound is part of the clip: Google DeepMind describes dialogue, effects and ambience created together with the picture. For Lite, Google writes that the output is video with sound, but it does not describe dialogue separately. So when speech is the heart of the scene, Veo 3.1 or Fast is the better choice.

Google does not confirm Hebrew in any of the versions. According to Google only English was evaluated, so it is worth writing the prompt in English. When you need a Hebrew voiceover, create it separately in a voice model whose official language list includes Hebrew, such as Eleven v3, and add it to the clip in editing. The voice models are compared in the guide Hebrew AI voiceover.

Start image and reference images

In all three versions you can upload a start image, which the clip begins from, and an end image, which it ends on. According to Google, in Veo 3.1 the transition between the start and the end image comes with sound too. This is how you animate an existing image, and the guide how to turn an image into a video explains it step by step.

Reference images are something else: you attach up to 3 images of a character, an object or a scene, and the model keeps their look in a new scene. Google documents them in Veo 3.1 and Fast, in an 8 second clip. It does not document them in Lite, but here you can pick up to 3 reference images in Lite too, in an 8 second clip, and in our test two reference images held through the whole clip. Right now an 8 second clip at 720p with reference images costs ₪15.00149.97 crVeo 3.1, 720p · 8s · with sound in Veo 3.1, ₪1.4114.06 crVeo 3.1 Fast, 720p · 8s · with sound in Fast and ₪0.707.03 crVeo 3.1 Lite, 720p · 8s · with sound in Lite.

Vertical video

All three versions make a landscape clip at 16:9 or a vertical one at 9:16, the format of Reels, Stories and Shorts. You choose the aspect ratio before you generate.

What to choose by use

If you are not sure, start with Lite or Fast to test the prompt, and move to Veo 3.1 only for the final version. All three versions work from the same balance and in the same studio, and before every generation the exact price is shown, which is the amount that leaves your balance.

If you are weighing Veo against another family, the guides Veo 3.1 vs Kling 3.0 and how to choose an AI video model compare them.

Questions and answers