All tasks / Video / Video generation
Best local AI for video generation
Open-weights models and tools that run entirely on your own hardware, ranked from public benchmarks. Local tools are rated by the best open model they run.
STop ranked: MiniMax H3 (MiniMax), 1,159 on Artificial Analysis Video Arena: Text to Video (AA-Video-T2V v2.1, with audio).13 options listed · data verified 19 Oct 2026
Ranked list
| Tier | Name | Approach | Runs | Rated by | Price | Skill | Compare |
|---|---|---|---|---|---|---|---|
| S |  MiniMax H3MiniMax | Models | Local | 1,159 on Artificial Analysis Video Arena: Text to Video (AA-Video-T2V v2.1, with audio) | Open weights | — | |
| S |  ComfyUIComfy Org | Workflow platforms | Local | via MiniMax H3 Cloud models: tier S | Free | advanced | |
| S |  WanGP (Wan2GP)deepbeepmeep (community project) | Apps | Local | via MiniMax H3 | Free | intermediate | |
| S |  SwarmUImcmonkeyprojects | Apps + Workflow platforms | Local | via MiniMax H3 | Free | intermediate | |
| A |  Cosmos3-Super-Image2VideoNVIDIA | Models | Local | 1,275 on Artificial Analysis Video Arena: Image to Video, no audio (AA-Video-I2V-Silent v1.0) | Open weights | — | |
| B |  MAGI-2 PreviewSand.ai | Models | Local | 1,045 on Artificial Analysis Video Arena: Text to Video (AA-Video-T2V v2.1, with audio) | ~640 GB VRAM | — | |
| C |  LTX-2.5Lightricks | Models | Local | 950 on Artificial Analysis Video Arena: Text to Video (AA-Video-T2V v2.1, with audio) | Open weights | — | |
| C |  HunyuanVideo 1.5Tencent | Models | Local | 1,169 on LMArena (Arena) Text-to-Video leaderboard | ~14 GB VRAM | — | |
| C |  Wan 2.2Alibaba (Wan-AI) | Models | Local | 1,132 on LMArena (Arena) Text-to-Video leaderboard | ~24 GB VRAM | — | |
| C |  SD.Nextvladmandic (community project) | Apps | Local | via LTX-2.5 | Free | intermediate | |
| C |  LTX DesktopLightricks | Apps | Local | via LTX-2.5 | Free | intermediate | |
| C |  Draw ThingsDraw Things | Apps | Local | via LTX-2.5 | Free | intermediate | |
| C |  AmuseTensorStack | Apps | Local | via Wan 2.2 | Free | beginner |
Tiers come only from public benchmarks (Artificial Analysis Video Arena: Text to Video (AA-Video-T2V v2.1, with audio), Artificial Analysis Video Arena: Image to Video, no audio (AA-Video-I2V-Silent v1.0), LMArena (Arena) Text-to-Video leaderboard); tools without a result of their own inherit the tier of the best model they offer. How we rank.