Mochi 1 (Genmo)
Source: https://github.com/genmoai/mochi
Structured fields
{
"archived": false,
"checked_at": "2026-10-17T00:00:00.000Z",
"usefulness": "10B-parameter open text-to-video model with a fully open video VAE and Apache-2.0 weights; LoRA fine-tuning support. Useful as a permissively licensed realistic-motion baseline, though the release is an explicit 480p research preview.",
"source_type": "repository",
"availability": "available",
"source_owner": "Genmo",
"next_check_at": "2026-10-20T00:00:00.000Z",
"hardware_notes": "Upstream: ~60 GB single-GPU in reference implementation; ~42 GB Diffusers; ~22 GB bf16 variant; <20 GB via ComfyUI.",
"known_limitations": [
"Newest news entry seen on the README is November 2024; no sign of newer model releases on the page.",
"Output 480p."
],
"upstream_updated_at": "2026-10-06T17:19:13.000Z"
}Evidence claims
- [upstream_claim] README describes Mochi 1 as a preview 10B AsymmDiT text-to-video model with AsymmVAE; single GPU needs ~60 GB VRAM, H100 recommended; ComfyUI runs it in under 20 GB. source
- [upstream_claim] Stated limitations: 480p only, minor warping on extreme motion, optimized for photorealism and weak on animated content. source
- [upstream_claim] Model card reports Diffusers example ~42 GB, bf16 variant ~22 GB with slight quality cost. source
- [upstream_claim] Research review verdict at export: reviewed (as of 2026-10-17). Recorded by the research pipeline, not a hub review outcome; evidence.review stays unreviewed until an independent reviewer records one.
Rights
| component | license | hosting | redistribution | modification | training | Evidence and notes |
|---|---|---|---|---|---|---|
| code | Apache-2.0 | permitted | permitted | permitted | unknown | Repository stated as Apache-2.0. |
| model_weights | Apache-2.0 | permitted | permitted | permitted | unknown | README says model released under Apache 2.0; HF metadata lists Apache-2.0. Card recommends additional safety measures before commercial deployment. |