Wan 2.2 5B
A solid pick for self-hosting and full data control.
Overview
Wan 2.2 5B is Alibaba's open-weight text-and-image-to-video model, generating 720p footage at 24fps through a highly compressed VAE. It trades the raw power of the family's 14B MoE variants for speed and accessibility, still carrying the cinematic aesthetic tuning of the Wan 2.2 line. Its limit: at 5B parameters, complex motion coherence lags behind the larger models. Apache 2.0 licensed and fully local-capable on a single consumer GPU like the RTX 4090, it keeps all data and compute on-premise.
Skill profile
Not disclosed
Strengths
- Open-source and self-hostable
Limitations
- API pricing not disclosed
Who is it for
- you want to control cost or self-host
- you have GDPR constraints
Ideal use cases
- Fast video prototyping
- On-device demo generation
- Image-to-video conversion
- Local creative experimentation
Access & availability
Key specifications
Privacy
Run it locally
| Quantization | Disk | RAM / VRAM | Typical hardware |
|---|---|---|---|
| Q4 · recommended | 2.8 GB | 5 GB | Any recent PC/Mac |
| Q8 · balanced | 5.4 GB | 8 GB | 16GB PC / M1+ Mac / 8GB GPU |
| FP16 · max quality | 10 GB | 13 GB | 32GB PC / 24GB Mac / RTX 4070 Ti+ |
Estimates for a moderate context. Long contexts need more RAM (KV cache).
Deploy
Copy-ready commands generated from this card. Adjust context length and GPU count to your hardware.
OpenAI-compatible server for production on NVIDIA GPUs.
pip install vllm
vllm serve Wan-AI/Wan2.2-TI2V-5B \
--max-model-len 32768 \
--tensor-parallel-size 1 \
--dtype autoAdvanced data · for expertsArchitecture, modalities, detailed cost, full benchmarks▾
| Arena Elo | — |
| MMLU | — |
| GPQA | — |
| HumanEval | — |
| SWE-Bench | — |
| MATH | — |