Optimization5 min read

Veo Pricing Math: Budget a Cinematic Trailer End to End

A 60 second trailer on Veo 3.1 at full quality lands around $96 in raw compute before retries. Here is the line by line math, including the retry buffer most teams forget.


You want a 60 second cinematic trailer. You are rendering on Veo 3.1 because Veo 4 is not on fal.ai yet. Here is what it actually costs.

The list price on Veo 3.1 is $0.40 per second at 1080p, with a hard 8 second per clip ceiling. Veo 3.1 Fast sits at roughly $0.25 for a 1080p 8 second clip, which works out to about $0.031 per second. Both tiers cap at 4K resolution and 8 seconds. Those are your two levers.

At 60 seconds total runtime, and 8 seconds per clip, you need a minimum of eight clips. In practice you want a little slack for the edit, so plan for ten clips. That gives the editor trims and reaction cuts without forcing reruns.

Line item breakdown of the trailer budget across shots and retries
Line item breakdown of the trailer budget across shots and retries

Ten clips at full quality, at 8 seconds each, is 80 seconds of billed video. At $0.40 per second, that is $32. Sounds cheap. It is not the real number.

The real number includes retries. On a trailer, where every shot has to land, expect a retry rate between 30% and 60% depending on how complex your prompts are. A simple locked off shot of a landscape might land on take one. A dolly in on a face with specific lighting might take three tries. Average it at 50% and you are rendering fifteen clips to get your ten keepers. Fifteen clips at 8 seconds at $0.40 is $48 for picture.

If the trailer has audio on the Veo side, because you are using the model's audio generation, there is no per second uplift on the current listed price. The $0.40 already includes audio when you pass generate_audio. If you are scoring externally and not using Veo's audio, nothing changes on cost.

Now add the exploration phase. Before you commit to your final prompts, you run tests. Shot finding, style locking, character sheet validation. Budget eight to twelve Fast renders for this, at roughly $0.25 each. Call it $3. Cheap insurance that saves you from burning full quality credits on a prompt that did not work.

So the running total is $48 for the picture run plus $3 for exploration, which is $51. For safety, add a 20% contingency for last minute reshoots the editor asks for. That is about $10. Round the whole thing to $60 to $65 for a clean 60 second trailer, rendered on Veo 3.1 at full 1080p quality.

Here is the call you would actually wire up for the main render loop.

JAVASCRIPT
1import { fal } from "@fal-ai/client";
2
3// or fal-ai/veo4/text-to-video once available
4async function renderShot(prompt) {
5 const result = await fal.subscribe("fal-ai/veo3.1/text-to-video", {
6 input: {
7 prompt,
8 aspect_ratio: "16:9",
9 duration: "8s",
10 resolution: "1080p",
11 generate_audio: true
12 },
13 logs: true
14 });
15 return result.data.video.url;
16}
17
18const shotList = [/* your ten prompts */];
19for (const p of shotList) {
20 const url = await renderShot(p);
21 console.log(url);
22}

One note. You can parallelize this with Promise.all, but if you are on a pay as you go tier, watch your rate limits. Running ten Veo calls at once is fine technically. It is also a way to burn $48 in ninety seconds if you realize after the fact that every prompt had a typo.

A two column cost comparison of full quality vs fast tier at 60 seconds
A two column cost comparison of full quality vs fast tier at 60 seconds

What about doing the whole thing on Veo 3.1 Fast? Ten clips plus a 50% retry rate is fifteen clips, at $0.25 each, which is $3.75. Add $3 exploration and $1 contingency and you are under $10 total. The catch is that Fast drops detail in the places a trailer cares about, which are faces, edges, and subtle motion. For a personal project or a social cut, Fast is fine. For a hero piece, use Fast for exploration and switch to full quality for the final renders. Split the budget and you get the best of both.

If you are planning against Veo 4, expect it to land in the same premium band as Veo 3.1 at launch, maybe with a quality tier that sits above the current $0.40 price point. Do not budget on a specific number. When the endpoint lands, rerun the math on the published price. The structure of the budget does not change. The per second cost might. Your retry rate might drop if Veo 4 is more consistent, which is the real upside. Dropping from 50% retries to 30% on the same ten clip plan is the difference between thirteen renders and fifteen, which is real money on a premium tier.

Final answer, current as of today. One 60 second cinematic trailer on Veo 3.1 at 1080p, with sensible retry budget and a small exploration phase, costs about $65. Not $32. Not $200. If someone quotes you either extreme, they either forgot retries or they are padding the number. Use $65 as your working figure and you will hit it within 15% on almost every real project.