# New video models: Wan 3, Seedance 2.5, Grok Imagine Lite

> Wan 3, Seedance 2.5 and Grok Imagine Video 1.5 Lite arrive on Tchavi: durations, resolutions, audio, price per second and a call you can copy.

- Published: 2026-10-18
- Author: Gael — Tchavi founder
- Canonical: https://tchavi.com/en/blog/new-video-models-wan-3-seedance-2-5-grok

---


Three video models join Tchavi: Wan 3.0, Seedance 2.5 and Grok Imagine Video 1.5 Lite. Wan 3.0 is for clips with ambient sound on a small budget, Seedance 2.5 for scenes where someone speaks, and Grok Lite for quick drafts from an image.

## In short

- Tchavi, an OpenAI-compatible AI API gateway you pay for with mobile money, adds `wan-3`, `seedance-2.5` and `grok-imagine-video-1.5-lite`.
- Wan 3.0: 2 to 15 s, 480p to 1080p, audio always included.
- Seedance 2.5: 4 to 15 s, 480p and 720p, lip-synced dialogue, including in French.
- Grok Imagine Video 1.5 Lite: 1 to 15 s, 480p to 1080p, no audio setting.

## Wan 3.0 (Alibaba)

Wan 3.0 generates video with sound: ambience and effects are included at no extra credits. Water, fabric and fast body movement are its strong point.

- Duration: 2 to 15 seconds.
- Resolutions: 480p, 720p, 1080p. Aspect ratios: 16:9, 9:16, 1:1, 4:3, 3:4.
- Inputs: a prompt, a first image, a `negative_prompt`, a `seed`.
- Limits: no last image; audio is always generated, with no off switch.

## Seedance 2.5 (ByteDance)

Seedance 2.5 writes the soundtrack together with the picture: a line you put in quotation marks is spoken and lip-synced. I checked it on a line in French for the demo on its page.

- Duration: 4 to 15 seconds.
- Resolutions: 480p and 720p (no 1080p). Aspect ratios: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9 and `adaptive`.
- Inputs: up to 9 reference images and 3 audio clips, first and last image, `seed`, `generate_audio`.
- Limit: reference videos are not accepted; Seedance 2.0 takes them.

## Grok Imagine Video 1.5 Lite (xAI)

The fast, light version of Grok Imagine Video. It animates a poster or a photo: our 5-second demo at 720p rendered in 29 seconds.

- Duration: 1 to 15 seconds.
- Resolutions: 480p, 720p, 1080p. Aspect ratios: 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3.
- Inputs: a prompt or a first image. No `seed`, last image or `negative_prompt`.
- Limits: each image you supply adds 22 credits; no audio switch (the clip in our test was almost silent).

## Prices

Credits first, then KES at October 2026 rates (1 credit is about KES 0.076):

| Model                       | Resolution          | Credits per second | KES per second   |
| --------------------------- | ------------------- | ------------------ | ---------------- |
| Wan 3.0                     | 480p / 720p / 1080p | 55 / 110 / 220     | 4.2 / 8.3 / 16.7 |
| Seedance 2.5                | 480p / 720p         | 230 / 508          | 17.4 / 38.5      |
| Grok Imagine Video 1.5 Lite | 480p / 720p / 1080p | 44 / 66 / 308      | 3.3 / 5.0 / 23.3 |

Audio is included in these prices: none of these models bills it separately. Example: 5 seconds of Wan 3.0 at 720p is 550 credits, about KES 42. With Grok Lite, 1080p costs almost five times 720p: run your tests at 720p.

These are public prices; some models are on discount right now, and each model page shows the price billed today. Wan 3.0 is one of them.

## One call to try it

Video generation is asynchronous: you submit the job, then poll it.

```bash
curl -X POST https://tchavi.com/api/v1/videos/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $TCHAVI_API_KEY" \
  -d '{
    "model": "wan-3",
    "prompt": "Wide shot at sunset on Diani beach, a fisherman runs past a colourful wooden boat",
    "duration": 5,
    "resolution": "720p"
  }'

# Then, with the returned id:
curl https://tchavi.com/api/v1/jobs/JOB_ID \
  -H "Authorization: Bearer $TCHAVI_API_KEY"
```

For Seedance 2.5, change `model` to `seedance-2.5` and write the spoken line in quotation marks in the prompt. If a model does not take a parameter, the model page and the [video documentation](https://tchavi.com/en/docs/video-generation) say so.

## How to try them

- Watch the demos: [Wan 3.0](https://tchavi.com/en/m/alibaba/wan-3), [Seedance 2.5](https://tchavi.com/en/m/bytedance/seedance-2.5), [Grok Imagine Video 1.5 Lite](https://tchavi.com/en/m/xai/grok-imagine-video-1.5-lite). Each page shows the prompt that produced the video.
- Read the [video generation documentation](https://tchavi.com/en/docs/video-generation) and the one on [jobs](https://tchavi.com/en/docs/jobs).

## FAQ

### Which model should I pick for a first test?

Wan 3.0 at 480p or 720p, with audio included. If you want a scene with a spoken line, try Seedance 2.5.

### Why does my clip from an image not have the aspect ratio I asked for?

With a first image, Wan 3.0 and Seedance 2.5 keep the image's framing: the `aspect_ratio` you send is ignored.

### What happens if a job fails?

The credits charged are refunded automatically if the job fails.

[create your account and top up KES 250](https://tchavi.com/en/register?next=%2Fdashboard%2Fbilling%3Famount%3D250%26currency%3DKES%26from%3Dblog-video-oct-en)

> _This article is part of the series "Building an AI product from Africa". Tchavi is an OpenAI-compatible AI API gateway: one key gives access to over 100 text, image, audio and video models, and you pay for it with mobile money, no bank card ([countries and operators](https://tchavi.com/en/docs/payments)). More guides on [the blog](https://tchavi.com/en/blog). Follow us: [LinkedIn](https://www.linkedin.com/company/tchavi) and [X](https://x.com/tchaviai)._