# Which AI video API should you use? Pick by task

Updated 2 Nov 2026

## Short answer

Pick the model by the job, then by price. For drafts use [Wan 2.6 Flash](/vidlab-api/models/wan-2.6-flash) or [Grok Imagine Video 1.5 Preview](/vidlab-api/models/grok-imagine-video-1.5-preview) at $0.03 per second. For social ads with sound use [Seedance 2.5](/vidlab-api/models/seedance-2.5), [Kling 3.0](/vidlab-api/models/kling-3.0) or [Grok Imagine Video 1.5](/vidlab-api/models/grok-imagine-video-1.5). For a talking head from a photo and a voice track use [OmniHuman 1.5](/vidlab-api/models/omnihuman-1.5) or [Kling Avatar Pro](/vidlab-api/models/kling-avatar-pro). For edits to an existing video use [Wan 2.7 Video Edit](/vidlab-api/models/wan-2.7-videoedit). All of them run through one VidLab API key, so switching models is a change of one string.

## Pick by task

Prices are US dollars per second of video at the setting shown, from the public catalogue. Each model's page lists exactly what it accepts.

| Task | Start with | Price per second |
| - | - | - |
| Cheap drafts and tests | [Wan 2.6 Flash](/vidlab-api/models/wan-2.6-flash) | $0.03 at 720p |
| Cheap drafts with sound included | [Grok Imagine Video 1.5 Preview](/vidlab-api/models/grok-imagine-video-1.5-preview) | $0.03 at 720p |
| Social ads, 9:16, with sound | [Seedance 2.5](/vidlab-api/models/seedance-2.5) | $0.21 at 720p |
| Social ads on a smaller budget | [Kling 3.0](/vidlab-api/models/kling-3.0) | $0.13 at 720p with audio |
| UGC and talking-head marketing video | [MiniMax H3 UGC](/vidlab-api/models/minimax-h3-ugc) | $0.05 at 768p |
| A photo plus a voice track, lip-synced | [OmniHuman 1.5](/vidlab-api/models/omnihuman-1.5) | $0.17 |
| The same, for longer audio | [Kling Avatar Pro](/vidlab-api/models/kling-avatar-pro) | $0.10 |
| Edit or restyle an existing video | [Wan 2.7 Video Edit](/vidlab-api/models/wan-2.7-videoedit) | $0.09 at 720p |
| Reference video as an input | [Kling 4.0](/vidlab-api/models/kling-4.0) | $0.11 at 720p |
| 4K output | [Veo 3.1 Lite](/vidlab-api/models/veo-3.1-lite) | $0.24 at 4K |
| Upscale finished video to 4K | [Topaz Video Upscale 4K](/vidlab-api/models/topaz-video-upscale) | $0.08 at 4K |

These are starting points chosen from what each model accepts and what it costs, not a ranking of quality. Quality depends on your prompt and your footage, which is why the last section is about testing.

## Decide in five questions

1. **What do you have?** Only text, a still image, a video, or a voice track? A lip-sync model needs an image and audio. An edit model needs a video. A plain text to video model needs neither.
2. **Does it need sound?** Some models make audio in the same pass and some sell it as an option. [Veo 3.1](/vidlab-api/models/veo-3.1) charges $0.20 per second without audio and $0.40 with it at 720p.
3. **How long is a clip?** Check the duration range. [Seedance 2.5](/vidlab-api/models/seedance-2.5) takes 4 to 30 seconds and [Veo 3.1](/vidlab-api/models/veo-3.1) takes 4, 6 or 8 seconds.
4. **Which resolution is the final?** Draft at 480p or 720p and render the final at 1080p. 4K is not worth paying for in a draft.
5. **What is the budget per usable clip?** Multiply the per-second price by the length, then by how many tries you expect. The [price comparison guide](/blog/cheapest-ai-video-generation-api) shows how.

## Test two models on the same prompt

The cheapest way to choose is to run the same prompt on two models at 480p or 720p and look. A 5 second 720p test is $0.15 on Wan 2.6 Flash and $1.05 on Seedance 2.5. With the API that is two calls that differ in one field:

```bash
for model in wan-2.6-flash seedance-2.5; do
  curl -s -X POST https://vidlab.ai/v1/generations \
    -H "Authorization: Bearer $VIDLAB_API_KEY" \
    -H "Content-Type: application/json" \
    -H "Idempotency-Key: $(uuidgen)" \
    -d "{\"model\":\"$model\",\"input\":{\"prompt\":\"A barista pours milk into a latte, close up\",\"resolution\":\"720p\",\"duration\":5,\"aspect_ratio\":\"9:16\"},\"max_price_usd\":1.10}"
done
```

A model refuses an option it does not accept instead of quietly changing it. Read the model's `params` from `GET /v1/models/{id}` and send only what it lists.

## Why use one API for all of them

Video models change every few weeks. A new model that suits your task should not mean a new account, a new SDK and a new invoice. On VidLab:

- One key and one request shape work for all 137 models, across video, image, audio, lip sync and upscaling.
- Every call can be priced first, for free, with `POST /v1/estimate`.
- Each call goes to the cheapest route that is up, and the price you pay is the catalogue price.
- A failed generation is not charged.
- Agents can use it too, through [MCP](/blog/vidlab-mcp-claude-cursor-codex) or the [skill file](/vidlab-api/SKILL.md).

## FAQ

### Which AI video model is best?

There is no single best model. It depends on the input you have, the length, the resolution, the sound and the look you want. Choose two that fit the task in the table and compare their output on your own prompt.

### Which is the cheapest AI video API?

The lowest 720p price in the table is $0.03 per second. The [guide to comparing per-second pricing](/blog/cheapest-ai-video-generation-api) explains what to check before you trust any price.

### Can I use more than one model in a project?

Yes. A common pattern is to make a first frame with an image model, animate it with a video model, and upscale the result. Every step is the same API with a different model id.

### How do I find a model's options?

`GET /v1/models/{id}` lists its params and every priced variant. The same page is on the [models page](/vidlab-api/models).

### Where do I start?

Create a key, run a free estimate for the model you are considering, and make one short test clip. Use the button below.