Modelverse
Search our library of open source models and deploy in seconds
Image Generation
FLUX.2-Klein-4B
Klein 4B is designed for workloads where latency matters more than maximum image fidelity.
FLUX.2-Klein-9B
Klein 9B balances image quality and low latency.
FLUX.1-Schnell
Schnell is built for speed and high volume.
FLUX.1-dev
FLUX.1 dev is for images where quality matters more than turnaround.
FLUX.2-dev
FLUX.2 dev is for detailed, high-fidelity images.
Qwen-Image
Qwen Image's standout strength is text rendering, producing clean typography and complex layouts in both English and Chinese.
Qwen-Image-2512-Lightning
Qwen Image 2512 Lightning is built for fast, prompt-driven text-to-image generation.
Qwen-Image-2.1
Qwen Image 2.1 is for images where the lettering and the small details must be right.
Z-Image-Turbo
Z-Image Turbo is for fast, photorealistic images.
Ideogram-4
Ideogram 4 is built for images that carry words.
HiDream-O1
HiDream O1 is for large, detailed images where fine texture matters.
Ideogram-4-Flash
Ideogram 4 Flash is built for images that carry words, when turnaround matters.
HiDream-O1-Flash
HiDream O1 Flash is for detailed, high-resolution images when you want them sooner.
Krea-2-Turbo
Krea 2 Turbo is for fast, high-resolution images.
Nano-Banana-2
Nano Banana 2 is designed for workloads where latency and cost matter more than maximum image fidelity.
Nano-Banana-Pro
Nano Banana Pro is well suited to prompt-driven generations that are held to a higher quality bar.
FLUX.2-Klein-4B-Edit
Klein 4B Edit is designed for editing workloads where latency matters more than maximum image fidelity.
FLUX.2-Klein-9B-Edit
Klein 9B Edit balances image quality and low latency.
Qwen-Image-Edit
Qwen Image Edit is suited to targeted single-image edits while keeping the rest of the frame intact.
Qwen-Image-Edit-2511-Lightning
Qwen Image Edit 2511 Lightning is built for instruction-based edits.
Qwen-Image-2.1-Edit
Qwen Image 2.1 Edit is for instruction-based edits that keep the subject recognizable.
Nano-Banana-2
Nano Banana 2 is designed for workloads where latency and cost matter more than maximum image fidelity.
Nano-Banana-2.1
In its editing mode, Nano Banana 2.1 takes your images as references and your prompt as the instruction.
Nano-Banana-Pro
Nano Banana Pro is for the same editing tasks as Nano Banana 2, but held to a higher quality bar.
Text-to-Video
LTX-2.3
LTX-2.3 is for short clips where you want the whole shot from a prompt.
LTX-2.5
LTX-2.5 is for short clips where you want the whole shot from a prompt.
MiniMax-H3-Fast
H3 generates the picture and the sound together, so the audio matches what happens on screen.
Veo-3.1
Veo 3.1 turns a written description straight into moving footage, with synchronised sound and no source image.
Wan-2.7
Wan 2.7 turns a written description into moving footage with sound, with one model and one set of controls.
Wan-3.0
Wan 3.0 turns a written description into moving footage with sound, and it holds a shot for longer than any earlier Wan generation.
Wan-3.0-Prime
Prime trades rate for turnaround: the same prompt, the same controls, less waiting.
HappyHorse-1.1
HappyHorse 1.1 turns a written description into moving footage with sound.
Image-to-Video
LTX-2.3
LTX-2.3 image-to-video is for bringing a still to life without losing it.
LTX-2.5
LTX-2.5 image-to-video is for bringing a still to life without losing it.
Veo-3.1
Veo 3.1 Image-to-Video keeps the composition of the still you give it, and lets the prompt drive the motion.
HappyHorse-1.0
HappyHorse 1.0 image-to-video starts the clip on the frame you give it, then adds motion and sound.
Wan-3.0
Wan 3.0 starts from a frame you already have and moves it, so the look of the first frame is settled before generation begins.
Wan-3.0-Prime
Prime starts from a frame you already have and moves it, with the same controls as the standard grade and less waiting.
HappyHorse-1.1
HappyHorse 1.1 keeps the look of the source frame and adds motion and sound to it.
Multi-Reference-to-Video
MiniMax-H3-Fast
H3 reads the references as a description of who or what is in the shot, not as the opening frame.
HappyHorse-1.0
HappyHorse 1.0 reference-to-video carries a subject from your images into a scene the prompt describes.
Wan-2.7
Wan 2.7 carries a subject you already have into new footage, instead of inventing one from the prompt.
Wan-3.0
Wan 3.0 carries subjects you already have into new footage, and takes more of them than any earlier Wan generation.
Wan-3.0-Prime
Prime carries subjects you already have into new footage, with the same reference ceiling as the standard grade and less waiting.
