Model
Trends
.ai
Model
Trends
.ai
the best open source ai models
Home
/
Wan Models

Wan AI models

Wan is Alibaba's open video generation family: Wan 2.1 and Wan 2.2 in 1.3B, 5B and 14B sizes, with text-to-video, image-to-video, first-and-last-frame and speech-to-video variants. Wan 2.2 introduced a mixture-of-experts design with separate high-noise and low-noise experts that produces smooth 720p motion with cinematic lighting on a single consumer GPU. The LoRAs listed here add camera moves, motion styles, characters and effects, and the lightning and distillation LoRAs that cut a render to four steps are among the most downloaded models on the site.

Hottest Wan models

instagirls
WAN
•
LORA
ultrareal
WAN
•
LORA
squish
WAN
•
LORA
high screen flood
WAN
•
LORA
live wallpaper
WAN
•
LORA
birdman pixelart1p5
WAN
•
LORA
ghibli style
WAN
•
LORA
intimate contact high
WAN
•
LORA
emotional breakdown high
WAN
•
LORA
gagging disgusted high
WAN
•
LORA
quality modifiers
WAN
•
LORA
hits the face
WAN
•
LORA

Wan guides

Errors & fixes
ComfyUI errors and how to fix them: red nodes, missing models, CUDA out of memory
The ComfyUI error messages people hit first, what each one means and the fix that works: missing custom nodes (red boxes), "Prompt outputs failed validation", CUDA out of memory, wrong model type in a loader, mat1 and mat2 shape errors, header deserialization, torch and xformers mismatches.
By Captain
2026-05-12
Models
What is a LoRA and how to use one in ComfyUI, Forge, Automatic1111 and on the cloud
LoRAs explained for people who want results, not maths: what a LoRA file changes, why it must match the base model's family, trigger words, weights, stacking several LoRAs, the folder and syntax for each UI, and how to try a LoRA on the cloud before downloading anything.
By Captain
2026-05-20
Errors & fixes
CUDA out of memory: how much VRAM each AI model needs and how to fit it
A VRAM table for SD 1.5, SDXL, Flux, FLUX.2, Qwen-Image, Z-Image, Wan 2.2 and HunyuanVideo, and the techniques that make a model fit: fp8 and GGUF quantisation, CPU offload, tiled VAE, resolution and frame limits, and when to stop fighting and run it hosted.
By Captain
2026-05-18
Cloud services
PirateDiffusion: run any model from Telegram, no GPU needed
The commands that matter in the PirateDiffusion Telegram bot: /render with model trigger words, (( )) and [[ ]] weighting, #recipes, /adetailer, /highdef and /facelift upscaling, /remix, /inpaint, ComfyUI workflows with /wf /run:, and the new // skills that pick the model for you.
By Captain
2026-10-03
Video models
Wan 2.2 video: text-to-video and image-to-video setup, settings and LoRAs
Getting first results from Wan 2.2: which model size to pick (5B or 14B), the high-noise / low-noise expert pair, resolution and frame counts that work, the lightning LoRAs that cut renders to four steps, prompt structure for motion, and how to run it hosted when the card is too small.
By Captain
2026-05-24
Apps
BitVector: Vector for Discord, Spyglass on the web and Prism chat
How to run hosted models and ComfyUI workflows from Discord with the Vector bot, from any browser with Spyglass, and by chatting with Prism: account linking, /ai render and /ai workflow, initial media for image-to-image and image-to-video, skills, and the fixes for the errors people hit first.
By Captain
2026-10-03
All guides
AD
No install required - run AI models on the cloud
No GPU, no downloads, no Python. These services run the same open source models on their own servers, so you can start generating from a browser, Discord or Telegram in a minute.

Other model families