Skip to content

Model catalog

Start with a task, then find a model and a working example.

First run? Start with one of these guides.

Wan2.1Start with 1.3B text-to-video Wan2.2Choose MoE, TI2V or Animate Qwen-Image-2.1Generate or edit images, standard or Turbo

Showing 18 / 18 model families

Video and audio

Wan2.1

Start with 1.3B text-to-video, or use I2V, VACE and distilled models.

  • Text-to-video
  • Image-to-video
  • VACE
Usage guide
Video and audio

Wan2.2

Video generation, speech-driven video and Animate task examples.

  • Text-to-video
  • Image-to-video
  • Animate
Usage guide
Video and audio

MiniMax-H3

Generate synchronized audio and video from text, first frames and reference inputs.

  • Text-to-audiovisual
  • Image-to-audiovisual
GitHub examples
Video and audio

LTX

Audio and video generation with LTX-2, 2.3 and 2.5, including distilled inference.

  • Text-to-audiovisual
  • Image-to-audiovisual
GitHub examples
Video and audio

HunyuanVideo-1.5

Environment, weights and inference instructions for HunyuanVideo-1.5.

  • Video generation
GitHub examples
Video and audio

Self-Forcing

Autoregressive video generation and corresponding FP8 inference configurations.

  • Autoregressive
  • Video generation
GitHub examples
Image generation and editing

Qwen-Image-2.1

Text-to-image generation and prompt-driven image editing, including Turbo.

  • Text-to-image
  • Image editing
  • Turbo
Usage guide
Image generation and editing

Qwen-Image

Qwen-Image, Edit and Layered models with their corresponding distilled weights.

  • Text-to-image
  • Image editing
  • Layered images
GitHub examples
Image generation and editing

Z-Image

Image generation, editing and multi-GPU examples for Z-Image-Turbo.

  • Text-to-image
  • Image-to-image
  • Turbo
GitHub examples
Image generation and editing

FLUX.2

Inference scripts and configurations for FLUX.2 dev and klein.

  • Image generation
  • dev
  • klein
GitHub examples
Image generation and editing

LongCat-Image

Text-to-image generation, image editing and Edit-Turbo configuration.

  • Text-to-image
  • Image editing
GitHub examples
Image generation and editing

BAGEL

Image generation and editing examples for BAGEL.

  • Image generation
  • Image editing
GitHub examples
Restoration and upscaling

SeedVR2

Image and video upscaling with 3B and 7B models.

  • Image upscaling
  • Video upscaling
GitHub examples
Restoration and upscaling

SwiftVR

Image and video restoration, upscaling and model conversion.

  • Image restoration
  • Video restoration
  • Upscaling
GitHub examples
Restoration and upscaling

RelayVSR

Entry points for the original 4× model and the native LightX2V 2× model.

  • Video upscaling
  • 2×
  • 4×
GitHub examples
Interaction and spatial tasks

EdenLive

Control video generation with text, a target first frame and proxy recordings.

  • Interactive video
  • Video-to-video
GitHub examples
Interaction and spatial tasks

WorldMirror

Single-step 3D reconstruction with depth, normals, cameras and spatial representations.

  • 3D reconstruction
  • Depth estimation
GitHub examples
Interaction and spatial tasks

WorldPlay

Inference entry points and configurations for WorldPlay.

  • World model
  • Interactive generation
GitHub examples

Need to prepare your environment? Start with the quickstart. See the full model ecosystem for more models and variants.

Wan2.1, Wan2.2 and Qwen-Image-2.1 have complete guides on this site. Other entries link to their repository instructions or scripts.