Image to video AI: animate any picture

Animate an image in Quad Studio

14 video models in Quad Chat accept a starting image, from ByteDance, Alibaba, Black Forest Labs, xAI, Google and MiniMax. Upload the frame in Quad Studio, describe the motion, and the clip lands in your library next to the source. 3 other video models are text-only and are listed on the text-to-video page. Video starts on the Ultra plan ($40/month) with Seedance 1 Pro Fast, Wan 3 Video, Grok Imagine Video 1.5 Lite and Seedance 2.0 Mini. Max ($200/month) adds Seedance 1.5 Pro, Wan 3 Video Prime, FLUX 3 Video, Grok Imagine Video, Grok Imagine Video 1.5, Veo 3.1 Fast and 4 more. The Free, Go and Pro plans do not include video.

Image to video models compared

Clip length is the shortest to the longest the model supports. Click a header to sort. Seedance 1 Pro Fast (Ultra, 2 to 12 seconds, up to 1080p); Seedance 1.5 Pro (Max, 4 to 12 seconds, up to 1080p, audio); Wan 3 Video (Ultra, 2 to 30 seconds, up to 1080p); Wan 3 Video Prime (Max, 2 to 30 seconds, up to 1080p); FLUX 3 Video (Max, 5 to 20 seconds, up to 720p, audio); Grok Imagine Video (Max, 1 to 15 seconds, up to 720p); Grok Imagine Video 1.5 (Max, 1 to 15 seconds, up to 1080p, audio); Veo 3.1 Fast (Max, 4 to 8 seconds, up to 1080p, audio); MiniMax H3 (Max, 4 to 15 seconds, up to 1440p, audio); MiniMax H3 Max (Max, 5 to 15 seconds, up to 768p); Veo 3.1 (Max, 4 to 8 seconds, up to 1080p, audio); Grok Imagine Video 1.5 Lite (Ultra, 1 to 15 seconds, up to 1080p, audio); Seedance 2.0 Mini (Ultra, 4 to 15 seconds, up to 720p, audio); Seedance 2.5 (Max, 4 to 30 seconds, up to 1080p, audio).

How to choose a video model

Straight from the catalog, not from quality benchmarks. Longest clips: Wan 3 Video, up to 30 seconds. The shortest supported clip across models is 1 second. Highest resolution: MiniMax H3, up to 1440p. Fast generation: Seedance 1 Pro Fast, Wan 3 Video Prime and MiniMax H3 Max. Native audio: Seedance 1.5 Pro, FLUX 3 Video, Grok Imagine Video 1.5, Veo 3.1 Fast, MiniMax H3 and 4 more. Turn sound on before you generate; clips with audio use more credits than silent ones. Closing frame: Seedance 1.5 Pro, Wan 3 Video, Wan 3 Video Prime, FLUX 3 Video, Veo 3.1 Fast, MiniMax H3 and 4 more also take a last image, so a clip can be steered from one image to another. Entry plan: Seedance 1 Pro Fast, Wan 3 Video, Grok Imagine Video 1.5 Lite and Seedance 2.0 Mini on Ultra ($40/month).

Credits, plans and privacy

Video shares one monthly credit balance with chat and images: Ultra 5,000 credits, Max 25,000. Longer, sharper clips cost more, and the estimate updates before you generate. Image and video models run on the Standard Lane, so the provider may retain prompts, reference images and results under its own data policy. Quad asks you to confirm this once before your first generation, and chat keeps the Private Lane on supported models.

Frequently asked questions

Which Quad Chat plan includes image-to-video generation?

Ultra ($40/month) is the first plan with Seedance 1 Pro Fast, Wan 3 Video, Grok Imagine Video 1.5 Lite and Seedance 2.0 Mini. Higher plans add Seedance 1.5 Pro, Wan 3 Video Prime, FLUX 3 Video, Grok Imagine Video, Grok Imagine Video 1.5 and 5 more. The Free, Go and Pro plans do not include video.

Which models can animate an image?

14 models from ByteDance, Alibaba, Black Forest Labs, xAI, Google and MiniMax, including Seedance 1 Pro Fast, Seedance 1.5 Pro, Wan 3 Video, Wan 3 Video Prime, FLUX 3 Video, Grok Imagine Video, Grok Imagine Video 1.5, Veo 3.1 Fast and 6 more. They sit next to chat and image models in one app.

How long can the clips be?

Length depends on the model, from 1 to 30 seconds (Wan 3 Video is the longest), at up to 1440p on MiniMax H3. Studio shows each model's supported lengths, resolutions and aspect ratios before you generate.

Can I make AI videos with sound?

Yes, on models with native audio: Seedance 1.5 Pro, FLUX 3 Video, Grok Imagine Video 1.5, Veo 3.1 Fast, MiniMax H3 and 4 more. Turn sound on before you generate; clips with audio use more credits than silent ones, and the estimate updates first.

What is a closing frame?

Some models also accept a last image, so a clip can be steered from one image to another: Seedance 1.5 Pro, Wan 3 Video, Wan 3 Video Prime, FLUX 3 Video, Veo 3.1 Fast, MiniMax H3 and 4 more. The rest animate from the first frame only.

Is my data private in Quad Studio?

Image and video models run on the Standard Lane, so the provider may retain prompts, reference images and results under its own data policy. Quad asks you to confirm this once before your first generation, and chat keeps the Private Lane on supported models.