The AI video and image models
inside Predis.ai

Pick the model that suits the shot. Every one runs from the same Predis.ai workspace, on the same credit balance.

Video models

Make video ads with any of these

Each clip below was made with that model. Choose one per video, or try the same prompt on two and keep the one you like.

Video

Veo 3.1

by Google

Makes the sound with the picture, so dialogue and background audio come out in the same clip. Suits product scenes and short spots with a voice.

See Google Veo 3.1
Video

Kling 3.0 Pro

by Kuaishou

Cuts between up to 6 shots in one clip while the subject stays the same. Suits short sequences in 1080p.

See Kling 3.0 Pro
Video

Kling 4.0

by Kuaishou

Makes up to 30 seconds in one pass, cutting between shots on its own, with stereo sound and dialogue. Suits a whole short ad in one clip.

See Kling 4.0
Video

Seedance 2.5

by ByteDance

Makes several shots in one clip and keeps the same subject across the cuts. Suits short sequences rather than a single shot.

See Seedance 2.5
Video

FLUX 3 Video

by Black Forest Labs

Makes up to 20 seconds in one take, with dialogue and sound made alongside the picture. Suits explainers and spots that need a voice.

See FLUX 3 Video
Video

Grok Imagine Video 1.5

by xAI

Brings a still image to life with sound, keeping the picture you started with intact. Suits posters, product shots and artwork you already have.

See Grok Imagine Video 1.5
Video

LTX 2.5 Pro

by LTX

Writes its own music and sound with the picture, so the action can follow the beat. Suits music-led clips at 1080p.

See LTX 2.5 Pro
Video edit

Runway Aleph 2.0

by Runway

Edits footage you already have from one sentence, keeping the motion and your original audio. Suits changing a setting, color or outfit without a reshoot.

See Runway Aleph 2.0
Video edit

Kling 4.0 Edit

by Kuaishou

Changes a video you already have from one sentence: the weather, the cast, the product or the whole style. Suits a reshoot you would rather not film.

See Kling 4.0 Edit
Video

Wan 2.7

by Alibaba

Fills the moment between a first and a last frame you choose. Suits reveals where the opening and the final shot both need to be exactly right.

See Wan 2.7
Video

MiniMax H3

by MiniMax

Comes in four versions, from Lite for quick drafts to H3 in 2K, all with sound. Suits testing an idea several times before you pick.

See MiniMax H3
More video models

See every video model in the app

These eleven are a start. Sign in to see the full list and try any of them on the same credits.

Open Predis.ai
Image models

Make image ads and posts with any of these

Each image below was made with that model. Some are strongest at text in the image, some at editing a photo you already have.

Made with GPT Image 2.5: a smoked-glass NOCTURNE perfume bottle on wet black marble Image

GPT Image 2.5

by OpenAI

Two versions: Flare for fast everyday images, Sunburst for edits where every detail has to hold.

See GPT Image 2.5
Made with GPT Image 2.6 Sunburst GIF: an animated GOOD MORNING mug with rising steam Image and GIF

GPT Image 2.6

by OpenAI

Makes animated GIFs as well as stills, with up to 16 frames at the speed you set. Suits banners, stickers and product loops.

See GPT Image 2.6
Made with Nano Banana Pro: a celadon KUMO tea tin with a crisp printed label Image

Nano Banana Pro

by Google

Clear text in many languages, translated labels, and one scene built from up to 14 references.

See Nano Banana Pro
Made with FLUX 3 Image: a VELO NORD e-bike ad on a misty forest road Image

FLUX 3 Image

by Black Forest Labs

Place each element of an ad with a box in the prompt, then change one thing and keep the rest.

See FLUX 3 Image
Made with Seedream 5 Pro: a lantern night market festival poster Image

Seedream 5 Pro

by ByteDance

Dense infographics, posters and app screens, with text in Chinese, English and more.

See Seedream 5 Pro
Made with Ideogram 4.5: a Lisbon café front after eight edits Image

Ideogram 4.5

by Ideogram

Edit the same image again and again, and the parts you didn't mention stay put.

See Ideogram 4.5
Made with Grok Imagine Image 2.0: a KOMET sneaker drop graphic Image

Grok Imagine Image 2.0

by xAI

One design laid out fresh for every shape, from a square post to a 20:9 banner.

See Grok Imagine Image 2.0
Made with Qwen Image 3: a broadsheet newspaper front page Image

Qwen Image 3

by Alibaba

One long prompt, a whole page back: front pages, menus and grids of infographics.

See Qwen Image 3
Made with MAI-Image-2.5: three EMBER & ASH hot sauce bottles on a shelf Image

MAI-Image-2.5

by Microsoft

Labels, packaging and signs, spelled the way you wrote them.

See MAI-Image-2.5
More image models

See every image model in the app

These nine are a start. Sign in to see the full list and try any of them on the same credits.

Open Predis.ai
One workspace

Every model, one credit balance

Switch models from one ad to the next without another subscription. Then edit, schedule and post from the same place.

  • 11 video models
  • 9 image models
  • 3-day trial, card required
Start Now