Is there a Qwen 2.1 video model?
No. The official release is named Qwen-Image-2.1, its Hugging Face repository is tagged text-to-image, and its model card describes text-to-image generation, image editing and transparent RGBA output. Video generation is not part of the model.
The search usually comes from two places. First, Alibaba's open video models are the Wan series, and Wan 2.1, published in February 2025 under Apache 2.0, shares the version number. Second, many people want to animate a picture they made with Qwen Image, which needs a separate image-to-video model.
| Qwen-Image-2.1 | Wan 2.1 | |
|---|---|---|
| Output | Still images | Video clips |
| Released | 20 September 2026 | February 2025 |
| License on Hugging Face | Qwen Research License | Apache 2.0 |
| Typical use | Posters, products, portraits, edits | Text-to-video and image-to-video |
How to turn a Qwen Image 2.1 picture into video
- Generate the still. On the homepage, select Qwen Image 2.1, write the scene and choose 16:9 or 9:16 to match the video format you want. Qwen Image 2.1 generation costs 1 credit per image.
- Download the image from the result or from your account history.
- Open the video workspace and choose an image-to-video model. The selector shows each model under its real name.
- Upload the still as the start image and describe only the motion: what moves, how the camera moves, and the pacing.
- Check the credit quote for the selected duration and resolution, then generate.
Videos made this way are outputs of the video model you select. They are not presented as Qwen-image-2.1 outputs.
Writing the motion prompt
The still already fixes subject, style and lighting, so the video prompt should not repeat them. Keep it to one continuous shot.
- Subject motion: "the cyclist rides slowly toward the camera", "steam rises from the cup".
- Camera motion: "slow push-in", "gentle orbit to the left", "locked-off tripod shot".
- Atmosphere: "fog drifts across the bridge", "rain streaks down the window".
- Constraints: "keep the lettering on the sign unchanged", "no cuts".
Slow camera orbit to the right around the bottle, ripples spread gently across the water, soft light stays constant, one continuous shot.Image-to-video models in this workspace
These are the image-to-video routes currently in the video selector. Each one is a separate model with its own name, clip lengths and resolutions; none of them is a Qwen model unless its name says so.
| Model | Clip lengths | Resolutions |
|---|---|---|
| H3 Max Turbo · First & Last Frame | 5s, 10s, 15s | 480p, 768p |
| MiniMax H3 Max · First & Last Frame | 5s, 10s, 15s | 480p, 768p |
| MiniMax H3 · First & Last Frame | 5s, 10s, 15s | 768p, 2K |
| Seedance 2.5 · Image to Video | 5s, 10s, 15s, 30s | 480p, 720p |
| Wan 3.0 · Image to Video | 5s, 10s, 15s, 20s, 25s, 30s | 480p, 720p, 1080p |
| MiniMax Hailuo 2.3 Pro · Image to Video | 5s | 1080p |
| MiniMax Hailuo 2.3 Standard · Image to Video | 6s, 10s | 768p |
| MiniMax Hailuo 2.3 Fast Pro · Image to Video | 6s | 1080p |
| MiniMax Hailuo 2.3 Fast · Image to Video | 6s, 10s | 768p |
| Seedance 2.0 Mini · Image to Video | 5s, 10s, 15s | 480p, 720p |
| Seedance 2.0 Fast · Image to Video | 5s, 10s, 15s | 480p, 720p |
| Seedance 2.0 Standard · Image to Video | 5s, 10s, 15s | 480p, 720p, 1080p |
| HappyHorse 1.1 · Image to Video | 5s, 10s, 15s | 720p, 1080p |
| Kling 3.0 Pro · Image to Video | 5s, 10s | 720p |
| FLUX 3 · Image to Video | 5s, 10s, 15s, 20s | 720p, 1080p |
The credit cost depends on the model, duration and resolution you choose. The exact quote is shown in the video workspace before you generate, and the pricing page lists every model.