RunAIHow.comLearn How To Run LocallyUpdated: OCTOBER 2026
Video Gen — Model Guides
Guides to today's video-generation models and services — capabilities, hardware demands, and commercial-use terms.

Sora (OpenAI)
Cloud onlyOpenAI
OpenAI's text-to-video model — long, coherent, physically-plausible clips with dialogue and sound.
4 min read · Open guide →

Google Veo (Gemini)
Cloud onlyGoogle DeepMind
Google's flagship video model inside Gemini — strong cinematography, native audio, and SynthID watermarking.
4 min read · Open guide →

Runway Gen-4
Cloud onlyRunway
The filmmaker's video generator — reference-driven consistency, camera control, and a professional toolset.
4 min read · Open guide →

Kling
Cloud onlyKuaishou
Kuaishou's video model with a generous free tier — strong motion physics and human animation.
4 min read · Open guide →

Wan 2.2 (Alibaba)
Open weightsAlibaba
The strongest open-weights video model — Apache-licensed text-to-video you can run on your own GPUs.
5 min read · Open guide →

Mochi 1 10B
Open weightsGenmo
Genmo's Apache-licensed 10B text-to-video model — smooth, coherent motion that a single 24 GB GPU can host.
4 min read · Open guide →

HunyuanVideo 13B
Open weightsTencent
Tencent's open 13B video model — cinema-grade motion quality that led the open-source rankings at release.
4 min read · Open guide →

LTX-Video 2B
Open weightsLightricks
The speed demon of open video — a 2B model that generates clips faster than they play back, on 8–12 GB cards.
4 min read · Open guide →
Also see the blog for deeper guides on hardware, quantization, and costs.