Every top AI image & video model, one studio

Flagship image model

Black Forest Labs’ flagship Flux release and the headline model here — the photorealism benchmark, running on credits in your browser.

In-context image editing

Tell it what to change and it changes only that — Black Forest Labs’ in-context editor in Pro and Max tiers.

Generate & edit images

Three generations of OpenAI image models in one place — from GPT-4o’s conversational editing to GPT Image 2’s 4K renders.

Photorealistic flagship

The highest-quality tier of Google’s Nano Banana line — built for images that pass as photographs.

Speed–quality sweet spot

The middle tier of Google’s Nano Banana family — polished results at a pace built for real production volume.

Viral image editor

The model that made “just tell it what to change” the way people edit photos — running at full speed in your browser.

Three-tier text to image

Google’s dedicated text-to-image engine, available in standard, Fast, and Ultra tiers from a single prompt box.

ByteDance flagship image model

The newest, most capable Seedream — ByteDance’s top text-to-image model, one prompt away in the Create studio.

Up to 4K output

ByteDance’s proven high-resolution image model — renders up to 4K with the realism the Seedream line is known for.

Proven 4K image model

The release that put Seedream on the map — dependable text-to-image up to 4K, still earning its keep.

Typography-strong images

The image model that actually spells — Ideogram V3 puts crisp, readable lettering into posters, logos, and layouts.

Alibaba’s image model

Alibaba’s Qwen team brings its prompt understanding to pictures — detailed text-to-image generation, one browser tab away.

Efficient text to image

A lean text-to-image model for high-volume everyday visuals — prompt in, picture out, credits stretched further.

Image finishing tools

The finishing pass for your generations — sharpen and enlarge with Crisp Upscale, or cut a subject free with Remove Background.

Cinema-grade video with audio

Google DeepMind’s flagship video model — cinematic motion, native audio, and frame-accurate control, ready to use in your browser.

Proven quality with audio

The workhorse Veo release — dependable cinematic quality with audio, keyframes, and full art-direction controls.

Budget Veo tier

All the Veo prompt-following you rely on, tuned for speed and the lowest credit cost in the family.

OpenAI’s viral video model

The model behind the internet’s most-shared AI clips — OpenAI’s Sora 2, available in your browser with no app invite required.

Character cameos from photos

Cast up to five recognizable characters into a Sora 2 scene from reference photos — cameo-style video without the Sora app.

Flagship Kling, up to 4K

Kuaishou’s flagship generation — longer shots, sharper detail up to 4K, and the lifelike human motion Kling is famous for.

Speed-tuned Pro quality

Pro-grade Kling output at turbo pace — the iteration workhorse of the family.

Pro, Standard & Master

The most versatile Kling lineup — three tiers, and the only one here with first-and-last-frame control.

Portrait to avatar video

One portrait in, a natural on-camera performance out — Kling’s dedicated avatar engine.

Video motion transfer

Record the move once, apply it to anyone — Kling’s motion-transfer engine puts real choreography on AI characters.

Up to 4K with references

ByteDance’s top Seedance release — 4K-capable renders, up to nine reference images, and an audio toggle, all inside your browser.

12-second clips, 8 ratios

ByteDance’s versatile mid-generation model — 12-second takes, the family’s widest aspect-ratio menu, and a lockable camera.

Pro & Lite value tiers

The original Seedance generation in five focused variants — pick Pro or Lite, text or image input, and generate in seconds.

Image-to-video specialist

MiniMax’s newest Hailuo release, purpose-built for turning a single still into fluid, believable motion.

Text & image to video

The all-rounder of the MiniMax lineup — prompt it with words or a picture and get physics-true motion either way.

Newest Wan release

Alibaba’s most capable Wan model yet — generate new shots, animate stills, edit existing footage, or compose scenes from up to nine reference images.

15-second clips

The long-form Wan: 15-second text-to-video with an audio toggle, plus image and video-to-video modes with quick Flash variants.

Dependable Wan workhorse

The steady mid-generation Wan — text and image to video with negative prompts for cleaner takes and fewer wasted credits.

Fastest Wan tier

The sketching tier of the Wan family — quick, inexpensive renders across text, image, and speech-driven video.

Motion transfer & swap

Point a performance at a character: transfer motion from real footage onto your subject, or replace the person in the frame outright.

Video & images, 30s clips

xAI’s all-in-one creative model — generate video up to 30 seconds and images from the same toolkit, then extend and upscale without leaving the studio.

Creative-studio video AI

The video model from the studio that helped invent AI filmmaking — prompt it or hand it a still, and it moves.

AI upscaling & enhancement

Topaz Labs’ enhancement engines, in your browser — the finishing pass that makes generated footage delivery-ready.

Photo + audio → talking video

Upload one portrait and one audio file — InfiniteTalk does the rest, mouth movement matched to every word.

ByteDance avatar model

ByteDance’s human-animation research, productized — one image of a person in, natural moving video out.

Re-dub any video

Keep the footage, change the words — VolcEngine regenerates the mouth so existing video speaks your new audio.

Flagship all-in-one model

Whatever you give it — a prompt, a frame, a stack of references, or a whole video — HappyHorse knows what to do with it.