AI Image Generator: Make a Picture, Then Turn It Into Video
Type what you want to see and Kensa draws it with Z-Image Turbo, Alibaba's open-weight 6B image model, in about 12–40 seconds for 1 credit per image. One click sends the picture to Image to Video, where Grok Image Video animates it from 5 credits (6 seconds, 480P). Accounts that sign up with Google get 15 free credits, valid for 72 hours.
By Alex Xu, founder of Kensa · Updated November 7, 2026
1 credit per image. Not charged if it fails. Usually 12–40 seconds.
Animals, products, places and illustrated characters animate best. Photo-real people may be refused by some video models; Seedance 2.0 accepts them.
How it works
- Step 1Describe the pictureWrite a sentence about the subject, setting, light and style, or start from an example. Text in quotes, such as a sign or a label, is drawn as written in English or Chinese.
- Step 2Generate it for 1 creditPick one of 7 aspect ratios (16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3). The image is ready in about 12–40 seconds. Failed or blocked images are not charged.
- Step 3Animate this imageThe image goes straight into Image to Video. Describe the motion, pick a model and generate; Grok Image Video starts at 5 credits for a 6-second clip with sound.
Z-Image Turbo on Kensa
| Model | Z-Image Turbo (Alibaba Tongyi-MAI, 6B parameters, Apache 2.0) |
|---|---|
| Price | 1 credit per image, refunded if it fails or is blocked |
| Aspect ratios | 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3 |
| Output size (measured) | 16:9 → 1376×768, 9:16 → 768×1376, 1:1 → 1536×1536, 3:4 → 896×1184, 3:2 → 1248×832 |
| Speed (measured) | 12–38 seconds per image on November 7, 2026 |
| Text in images | English and Chinese; requested text was spelled right in all 6 test prompts (one character came out in a variant form) |
| Safety | Prompts checked before generation, images checked after |
What we saw in testing
We ran six prompts through the same API before launch: bilingual shop signs, a Chinese menu, a poster, a product shot, a cat and an anime character. All the English and Chinese text we asked for was legible and spelled right; one Chinese character (单) came out in its Japanese form (単), and small background text we did not ask for was gibberish.
We then animated the cat picture with Grok Image Video: a 6-second clip with sound, ready in about a minute. The cat blinked and the rain ran down the glass, but it also turned its whole body and its fur pattern changed. Ask for small motions, such as a blink, a head turn or light moving across the scene, to keep the subject consistent.
Full results, local setup and API prices are in our Z-Image Turbo test write-up. Already have a photo? Go straight to Image to Video.
FAQ
- How much does an image cost?
- 1 credit per image, at any of the 7 aspect ratios. If generation fails or the image is blocked by the safety filter, the credit is returned automatically.
- Can I try it for free?
- Yes. Accounts that sign up with Google get 15 free credits, valid for 72 hours. That covers an image (1 credit) plus two Grok Image Video clips (5 credits each).
- Which model makes the images?
- Z-Image Turbo, an open-weight 6B text-to-image model from Alibaba's Tongyi-MAI team, released under Apache 2.0. Kensa runs it through a hosted API. Kensa is an independent product and is not affiliated with Alibaba.
- Why does it take 12–40 seconds if Z-Image Turbo is called sub-second?
- The sub-second figure is for a dedicated H800 GPU. Kensa calls a shared API, and in our November 7, 2026 test each image took 12–38 seconds end to end. We also save the image to Kensa's storage and run a safety check before showing it.
- Can I animate images of people?
- You can try, but several video models refuse photo-realistic people, so the video step may be blocked. Seedance 2.0 accepts realistic people. Animals, products, places and illustrated characters work with every model.
- What is not allowed?
- Sexual or violent content, real public figures and copyrighted characters are not allowed, and prompts that ask for them are blocked. Prompts are checked before generation and images after it, as described in the Acceptable Use Policy.
- Where are my images stored?
- Each image is saved to Kensa's file storage under a random, unlisted link so you can open it and use it in a video; anyone who has that link can open it. Your most recent images are also listed in your browser on this page.
Kensa is an independent product and is not affiliated with Alibaba or the Tongyi-MAI team. Z-Image Turbo is developed by Tongyi-MAI and is offered here as an integration. See the Acceptable Use Policy and pricing.