MAGI-1 alternatives

MAGI-1 is an open-source model, Free (open source). These 12 tools are in the same category and share the most vendor-stated features with it.

Why each one is here

• CogVideoX: Open-weight text-to-video and image-to-video models (CogVideoX-2B, 5B and 1.5) that run on desktop GPUs; CogVideoX1.5 makes 10-second clips at 1360x768. Shared with MAGI-1: text to video, image to video, open weights.
• HunyuanVideo: Tencent's open-weight video generation model with text-to-video and image-to-video variants, single- and multi-GPU inference. Shared with MAGI-1: text to video, image to video, open weights.
• LTX-2 / LTX-Video: Lightricks' open-weight audio-video model family: LTX-2.5 generates synchronized audio and video, with a UHD 4K mode; earlier LTX-Video distilled models make up to 60-second clips. Shared with MAGI-1: text to video, image to video, open weights.
• Open-Sora: Apache-2.0 open-source video generation project from HPC-AI Tech (Open-Sora 2.0), unrelated to OpenAI's Sora. Shared with MAGI-1: text to video, image to video, open weights.
• SkyReels: Skywork's open-weight video models: SkyReels-V2 uses autoregressive diffusion forcing for long ("infinite-length") generation at 540p/720p, with V3 weights released in 2026. Shared with MAGI-1: image to video, extend video, open weights.
• Wan: Alibaba's open-weight video models (Wan 2.x): text-to-video and image-to-video at 720p/24fps that run on a single consumer GPU such as an RTX 4090. Shared with MAGI-1: text to video, image to video, open weights.
• Genmo: Research lab behind Mochi 1, an open-source text-to-video model you can run yourself or in ComfyUI, plus a hosted playground. Shared with MAGI-1: text to video, open weights.
• Luma Dream Machine: Luma's creative platform (Dream Machine / Luma Agents) that runs Luma and third-party image and video models on a shared credit balance. Shared with MAGI-1: text to video, image to video, extend video.
• Magic Hour: All-in-one AI video, image and audio app (face swap, lip sync, image-to-video, templates) with API access on paid plans. Shared with MAGI-1: text to video, image to video, extend video.
• PixVerse: Text- and image-to-video generator with its own PixVerse models, templates and an app, popular for short social clips. Shared with MAGI-1: text to video, image to video, extend video.
• Pollo AI: Credit-based AI video and image platform with its own Pollo models plus third-party models such as Kling and Seedance. Shared with MAGI-1: text to video, image to video, extend video.
• Stable Video Diffusion: Stability AI's image-to-video latent diffusion model (SVD-XT: 25 frames at 576x1024), downloadable under the Stability AI Community License. Shared with MAGI-1: image to video, open weights.

Head-to-head comparisons

Lists that include MAGI-1

How we pick alternatives: same product group (generators, avatar tools, or editors and clip tools), then ranked by shared vendor-stated features and published pricing. Methodology.