wavespeed/wan2.2
20
1<!doctype html>2<html lang="en">3<head>4 <meta charset="utf-8" />5 <meta name="viewport" content="width=device-width, initial-scale=1" />6 <title>Wan 2.2 — open large-scale video generative models</title>7 <meta name="description" content="Reference for Wan 2.2, the open video generation model family from Alibaba's Tongyi Lab: architecture, released checkpoints, and hosted API usage." />8 <link rel="canonical" href="https://wavespeed.ai/collections/wan-2-2?utm_source=huggingface&utm_medium=space&utm_campaign=wan2_2" />9 10 <meta property="og:type" content="website" />11 <meta property="og:title" content="Wan 2.2 — open large-scale video generative models" />12 <meta property="og:description" content="Reference for Wan 2.2, the open video generation model family from Alibaba's Tongyi Lab: architecture, released checkpoints, and hosted API usage." />13 <meta property="og:url" content="https://wavespeed.ai/collections/wan-2-2" />14 <meta name="twitter:card" content="summary_large_image" />15 16 <link rel="stylesheet" href="style.css" />17</head>18<body>19 <header class="site-header">20 <div class="wrap">21 <a class="brand" href="https://wavespeed.ai?utm_source=huggingface&utm_medium=space&utm_campaign=wan2_2" target="_blank" rel="noopener">22 <span class="brand-mark" aria-hidden="true"></span>23 <span>WaveSpeed AI</span>24 </a>25 <nav class="header-nav">26 <a class="jump opt" href="#architecture">Architecture</a>27 <a class="jump opt" href="#checkpoints">Checkpoints</a>28 <a class="jump" href="#run">Run it</a>29 <a class="jump" href="#resources">Resources</a>30 <a href="https://wavespeed.ai/collections/wan-2-2?utm_source=huggingface&utm_medium=space&utm_campaign=wan2_2" target="_blank" rel="noopener">wavespeed.ai ↗</a>31 </nav>32 </div>33 </header>34 35 <main class="wrap">36 <div class="hero">37 <p class="eyebrow">Alibaba · Tongyi Lab</p>38 <h1>Wan 2.2</h1>39 <p class="lede">An open family of large-scale video generative models covering text-to-video, image-to-video and a compact hybrid text/image-to-video checkpoint. Weights are published on Hugging Face under Apache-2.0.</p>40 <ul class="meta">41 <li><b>Developer</b> Alibaba Tongyi Lab</li>42 <li><b>Task</b> text-to-video · image-to-video</li>43 <li><b>License</b> Apache-2.0</li>44 <li><b>Released</b> July 2025</li>45 </ul>46 </div>47 48 <section id="architecture">49 <h2>What changed in 2.2</h2>50 <p class="section-note">Wan 2.2 revises the 2.1 architecture in three places. The claims below are the authors' own, taken from the model cards and technical report.</p>51 <div class="grid">52 <div class="card">53 <h3>Mixture-of-experts denoiser</h3>54 <p>The denoising trajectory is split across specialised expert models rather than one monolithic network, which raises total parameter count without a matching rise in per-step inference cost.</p>55 </div>56 <div class="card">57 <h3>Curated aesthetic supervision</h3>58 <p>Training data carries explicit labels for lighting, composition, contrast and colour tone, so cinematographic attributes can be steered from the prompt instead of emerging by chance.</p>59 </div>60 <div class="card">61 <h3>Larger motion corpus</h3>62 <p>The authors report training on 65.6% more images and 83.2% more video than Wan 2.1, aimed primarily at motion fidelity and prompt adherence.</p>63 </div>64 <div class="card">65 <h3>High-compression VAE</h3>66 <p>The TI2V-5B checkpoint pairs with a Wan2.2-VAE at a 16×16×4 compression ratio, which is what makes 720p/24fps generation practical at that model size.</p>67 </div>68 </div>69 </section>70 <section id="checkpoints">71 <h2>Released checkpoints</h2>72 <p class="section-note">All weights are on the Hugging Face Hub under the Wan-AI organisation.</p>73 <div class="table-scroll">74 <table>75 <thead><tr><th>Checkpoint</th><th>Task</th><th>Params</th><th>Weights</th></tr></thead>76 <tbody>77 <tr><td><code>Wan2.2-T2V-A14B</code></td><td>Text-to-video</td><td>14B (MoE)</td><td><a href="https://huggingface.co/Wan-AI/Wan2.2-T2V-A14B" target="_blank" rel="noopener">base</a> · <a href="https://huggingface.co/Wan-AI/Wan2.2-T2V-A14B-Diffusers" target="_blank" rel="noopener">diffusers</a></td></tr>78 <tr><td><code>Wan2.2-I2V-A14B</code></td><td>Image-to-video</td><td>14B (MoE)</td><td><a href="https://huggingface.co/Wan-AI/Wan2.2-I2V-A14B" target="_blank" rel="noopener">base</a> · <a href="https://huggingface.co/Wan-AI/Wan2.2-I2V-A14B-Diffusers" target="_blank" rel="noopener">diffusers</a></td></tr>79 <tr><td><code>Wan2.2-TI2V-5B</code></td><td>Text + image-to-video, 720p/24fps</td><td>5B</td><td><a href="https://huggingface.co/Wan-AI/Wan2.2-TI2V-5B" target="_blank" rel="noopener">base</a> · <a href="https://huggingface.co/Wan-AI/Wan2.2-TI2V-5B-Diffusers" target="_blank" rel="noopener">diffusers</a></td></tr>80 </tbody>81 </table>82 </div>83 </section>84 <section id="run">85 <h2>Run it</h2>86 <p class="section-note">If you would rather not provision GPUs, the same checkpoints are served as a hosted endpoint. Available variants: <code>wan-2.2/t2v-480p</code>, <code>t2v-720p</code>, <code>i2v-480p</code> and <code>i2v-720p</code>.</p>87 <div class="code">88 <div class="code-tabs" role="tablist">89 <button type="button" role="tab" aria-selected="true" data-panel="run-0">cURL</button>90 <button type="button" role="tab" aria-selected="false" data-panel="run-1">Python</button>91 <button type="button" role="tab" aria-selected="false" data-panel="run-2">JavaScript</button>92 </div>93 <pre id="run-0" role="tabpanel"><code># 1. submit the job94curl -X POST "https://api.wavespeed.ai/api/v3/wavespeed-ai/wan-2.2/t2v-480p" \95 -H "Authorization: Bearer $WAVESPEED_API_KEY" \96 -H "Content-Type: application/json" \97 -d '{98 "prompt": "A paper boat drifting down a rain-slicked gutter at dusk, shallow depth of field, warm street lights",99 "duration": 5,100 "enable_sync_mode": false101 }'102 103# -> {"code": 200, "data": {"id": "<request-id>", "status": "created", ...}}104 105# 2. poll until status is "completed"106curl "https://api.wavespeed.ai/api/v3/predictions/<request-id>/result" \107 -H "Authorization: Bearer $WAVESPEED_API_KEY"108 109# -> {"code": 200, "data": {"status": "completed", "outputs": ["https://..."]}}</code></pre>110 <pre id="run-1" role="tabpanel" hidden><code>import os, time, requests111 112API = "https://api.wavespeed.ai/api/v3"113KEY = os.environ["WAVESPEED_API_KEY"]114HEADERS = {"Authorization": f"Bearer {KEY}"}115 116# submit117res = requests.post(118 f"{API}/wavespeed-ai/wan-2.2/t2v-480p",119 headers={**HEADERS, "Content-Type": "application/json"},120 json={121 "prompt": "A paper boat drifting down a rain-slicked gutter at dusk, shallow depth of field, warm street lights",122 "duration": 5,123 "enable_sync_mode": false124 },125 timeout=30,126)127res.raise_for_status()128request_id = res.json()["data"]["id"]129 130# poll131while True:132 data = requests.get(133 f"{API}/predictions/{request_id}/result",134 headers=HEADERS,135 timeout=30,136 ).json()["data"]137 138 if data["status"] == "completed":139 print(data["outputs"][0])140 break141 if data["status"] == "failed":142 raise RuntimeError(data.get("error", "generation failed"))143 time.sleep(1.5)</code></pre>144 <pre id="run-2" role="tabpanel" hidden><code>const API = "https://api.wavespeed.ai/api/v3";145const KEY = process.env.WAVESPEED_API_KEY;146const headers = { Authorization: `Bearer ${KEY}` };147 148// submit149const submit = await fetch(`${API}/wavespeed-ai/wan-2.2/t2v-480p`, {150 method: "POST",151 headers: { ...headers, "Content-Type": "application/json" },152 body: JSON.stringify({153 "prompt": "A paper boat drifting down a rain-slicked gutter at dusk, shallow depth of field, warm street lights",154 "duration": 5,155 "enable_sync_mode": false156 }),157});158const { data: { id } } = await submit.json();159 160// poll161for (;;) {162 const res = await fetch(`${API}/predictions/${id}/result`, { headers });163 const { data } = await res.json();164 165 if (data.status === "completed") {166 console.log(data.outputs[0]);167 break;168 }169 if (data.status === "failed") throw new Error(data.error ?? "generation failed");170 await new Promise((r) => setTimeout(r, 1500));171}</code></pre>172 </div>173 <div class="callout"><p>Requests are asynchronous: <code>POST</code> returns a request id, then you poll <code>/predictions/<id>/result</code> until <code>status</code> is <code>completed</code>. Set <code>enable_sync_mode: true</code> to have the call block and return outputs directly.</p><p>API keys are created in the <a href="https://wavespeed.ai/dashboard?utm_source=huggingface&utm_medium=space&utm_campaign=wan2_2" target="_blank" rel="noopener">WaveSpeed dashboard</a>.</p></div>174 <div class="btn-row">175 <a class="btn" href="https://wavespeed.ai/collections/wan-2-2?utm_source=huggingface&utm_medium=space&utm_campaign=wan2_2" target="_blank" rel="noopener">Open Wan 2.2 on WaveSpeed</a>176 <a class="btn secondary" href="https://wavespeed.ai/docs?utm_source=huggingface&utm_medium=space&utm_campaign=wan2_2" target="_blank" rel="noopener">API reference</a>177 </div>178 </section>179 <section id="resources">180 <h2>Resources</h2>181 <ul class="links">182 <li><a href="https://huggingface.co/Wan-AI" target="_blank" rel="noopener"><span>Wan-AI on Hugging Face</span><span class="host">huggingface.co</span></a></li>183 <li><a href="https://github.com/Wan-Video/Wan2.2" target="_blank" rel="noopener"><span>Wan2.2 on GitHub</span><span class="host">github.com</span></a></li>184 <li><a href="https://wavespeed.ai/collections/wan-2-2?utm_source=huggingface&utm_medium=space&utm_campaign=wan2_2" target="_blank" rel="noopener"><span>Hosted endpoints</span><span class="host">wavespeed.ai</span></a></li>185 <li><a href="https://wavespeed.ai/docs?utm_source=huggingface&utm_medium=space&utm_campaign=wan2_2" target="_blank" rel="noopener"><span>WaveSpeed API docs</span><span class="host">wavespeed.ai</span></a></li>186 </ul>187 </section>188 </main>189 190 <footer class="site-footer">191 <div class="wrap">192 <p>This page is a model reference maintained by WaveSpeed AI. The model itself is developed and released by its respective authors; trademarks belong to them. WaveSpeed AI provides hosted inference for it.</p>193 <div class="footer-links">194 <a href="https://wavespeed.ai?utm_source=huggingface&utm_medium=space&utm_campaign=wan2_2" target="_blank" rel="noopener">WaveSpeed AI</a>195 <a href="https://wavespeed.ai/docs?utm_source=huggingface&utm_medium=space&utm_campaign=wan2_2" target="_blank" rel="noopener">Docs</a>196 <a href="https://huggingface.co/wavespeed" target="_blank" rel="noopener">Hugging Face</a>197 </div>198 </div>199 </footer>200 201 <script src="tabs.js"></script>202</body>203</html>204