Top Open Source Text-to-Video Models 2026: Run Locally for Free

Free open source text-to-video models you can run locally with no cloud costs.

SVD-4D from Stability AI has 48K GitHub stars. AnimateDiff v3 has 32K stars. CogVideoX from Tsinghua has 18K stars. Open-Sora has 24K stars. All run on consumer GPUs with 8-24GB VRAM. Full privacy and zero API fees.