Files read on Hugging Face: 2026-10-24

Mochi 1 preview on Mac: files, RAM and published times

Mochi 1 preview is a 10B-parameter text-to-video model from Genmo built on an Asymmetric Diffusion Transformer. It generates 480p video and is a research preview, and one complete Mac measurement exists, on a 64 GB M4 Max.

Genmo · Updated 2026-10-24
Answer

Mochi 1 preview runs on a Mac with 64 GB or more. A published run on M4 Max · 64 GB used the Official Diffusers BF16 variant build (source, 2026-09-02).

Smallest Mac with a published run
64 GBM4 Max · 64 GB · Official Diffusers BF16 variant
Smallest complete download
7.67 GBcalcuis GGUF Q3_K_M with Q4_0 T5 and FP8 VAE
Files read
2026-10-24Hugging Face

Apache License 2.0. Commercial use: allowed. Excluded territories: none listed.

See the exact GGUF files and sizes

Table 1

What do you download?

Only builds with every required component and an exact probed file size are shown.

calcuis GGUF Q3_K_M with Q4_0 T5 and FP8 VAE

gguf · Q3_K_M · 7.67 GB complete

Mac routes: ComfyUI

Exact component files for calcuis GGUF Q3_K_M with Q4_0 T5 and FP8 VAE
FileComponentRequiredDownloadSource
mochi-q3_k_m.ggufcalcuis/mochiAsymmDiTditYes4.31 GBHugging FaceRead 2026-10-24
t5xxl_fp16-q4_0.ggufcalcuis/mochiT5-XXL text encodertext_encoderYes2.90 GBHugging FaceRead 2026-10-24
mochi_vae_fp8_e4m3fn.safetensorscalcuis/mochiAsymmVAEvaeYes0.46 GBHugging FaceRead 2026-10-24

Build sources: calcuis mochi GGUF card.

Comfy-Org repackaged FP8 scaled

safetensors · FP8 scaled · 16.10 GB complete

Mac routes: ComfyUI

Exact component files for Comfy-Org repackaged FP8 scaled
FileComponentRequiredDownloadSource
split_files/diffusion_models/mochi_preview_fp8_scaled.safetensorsComfy-Org/Mochi-PreviewAsymmDiTditYes10.03 GBHugging FaceRead 2026-10-24
split_files/text_encoders/t5xxl_fp8_e4m3fn_scaled.safetensorsComfy-Org/Mochi-PreviewT5-XXL text encodertext_encoderYes5.16 GBHugging FaceRead 2026-10-24
split_files/vae/mochi_vae.safetensorsComfy-Org/Mochi-PreviewAsymmVAEvaeYes0.92 GBHugging FaceRead 2026-10-24

Build sources: Comfy-Org Mochi Preview repackaged card.

Comfy-Org repackaged BF16 with FP16 T5

safetensors · BF16 · 30.76 GB complete

Mac routes: ComfyUI

Exact component files for Comfy-Org repackaged BF16 with FP16 T5
FileComponentRequiredDownloadSource
split_files/diffusion_models/mochi_preview_bf16.safetensorsComfy-Org/Mochi-PreviewAsymmDiTditYes20.06 GBHugging FaceRead 2026-10-24
split_files/text_encoders/t5xxl_fp16.safetensorsComfy-Org/Mochi-PreviewT5-XXL text encodertext_encoderYes9.79 GBHugging FaceRead 2026-10-24
split_files/vae/mochi_vae.safetensorsComfy-Org/Mochi-PreviewAsymmVAEvaeYes0.92 GBHugging FaceRead 2026-10-24

Build sources: Comfy-Org Mochi Preview repackaged card.

Official Diffusers BF16 variant

safetensors · BF16 · 40.02 GB complete

Mac routes: Diffusers

Exact component files for Official Diffusers BF16 variant
FileComponentRequiredDownloadSource
transformer/diffusion_pytorch_model.bf16-00001-of-00003.safetensorsgenmo/mochi-1-previewAsymmDiTditYes9.94 GBHugging FaceRead 2026-10-24
transformer/diffusion_pytorch_model.bf16-00002-of-00003.safetensorsgenmo/mochi-1-previewAsymmDiTditYes9.93 GBHugging FaceRead 2026-10-24
transformer/diffusion_pytorch_model.bf16-00003-of-00003.safetensorsgenmo/mochi-1-previewAsymmDiTditYes0.19 GBHugging FaceRead 2026-10-24
text_encoder/model-00001-of-00004.safetensorsgenmo/mochi-1-previewT5-XXL text encodertext_encoderYes4.99 GBHugging FaceRead 2026-10-24
text_encoder/model-00002-of-00004.safetensorsgenmo/mochi-1-previewT5-XXL text encodertext_encoderYes5.00 GBHugging FaceRead 2026-10-24
text_encoder/model-00003-of-00004.safetensorsgenmo/mochi-1-previewT5-XXL text encodertext_encoderYes4.87 GBHugging FaceRead 2026-10-24
text_encoder/model-00004-of-00004.safetensorsgenmo/mochi-1-previewT5-XXL text encodertext_encoderYes4.19 GBHugging FaceRead 2026-10-24
vae/diffusion_pytorch_model.bf16.safetensorsgenmo/mochi-1-previewAsymmVAEvaeYes0.92 GBHugging FaceRead 2026-10-24

Build sources: Mochi 1 model card.

Table 2

Does it fit your Mac's RAM?

Published names the exact measured machine. est. is a memory calculation. unknown means the run verdict is not established, not that the model fails.

Run verdicts by Mac memory tier
Mac RAMComplete buildDownloadRun verdictBasis
8 GBcalcuis GGUF Q3_K_M with Q4_0 T5 and FP8 VAEQ3_K_M · ComfyUI7.67 GBunknownunknown
16 GBcalcuis GGUF Q3_K_M with Q4_0 T5 and FP8 VAEQ3_K_M · ComfyUI7.67 GBunknownunknown
24 GBcalcuis GGUF Q3_K_M with Q4_0 T5 and FP8 VAEQ3_K_M · ComfyUI7.67 GBunknownunknown
32 GBcalcuis GGUF Q3_K_M with Q4_0 T5 and FP8 VAEQ3_K_M · ComfyUI7.67 GBunknownunknown
48 GBcalcuis GGUF Q3_K_M with Q4_0 T5 and FP8 VAEQ3_K_M · ComfyUI7.67 GBunknownunknown
64 GBOfficial Diffusers BF16 variantBF16 · Diffusers40.02 GBRuns · PublishedPublishedMacBook Pro · Apple M4 Max · 64 GB RAM
96 GBOfficial Diffusers BF16 variantBF16 · Diffusers40.02 GBRuns · Published on M4 Max · 64 GBPublishedMacBook Pro · Apple M4 Max · 64 GB RAM
128 GBOfficial Diffusers BF16 variantBF16 · Diffusers40.02 GBRuns · Published on M4 Max · 64 GBPublishedMacBook Pro · Apple M4 Max · 64 GB RAM

How does it rank in blind votes?

#48 of 48 in text to video blind votes, from 5,893 votes.

Quality: blind human votes from mochi-v1 (arena.ai), CC BY 4.0, snapshot 2026-09-22. Scores are for the Arena checkpoint, not your local quantized build.

Table 3

What published Mac times are available?

Third-party results keep their original machine and settings. ModelFit does not run these benchmarks.

Published Mac generation measurements
Machine and RAMSettingsAppResultSource
MacBook Pro · Apple M4 Max · 64 GB RAM848x480 · 2.8 s · 64 steps · 84 frames · BF16 · Diffusers MochiPipeline bf16 variant on PyTorch MPS, seed 42, one run. Swap of 19.8 GB reported. Peak is sampled every 0.5 s. Sampler, CFG and attention implementation are not reported. The date is the article date; the run date is not stated.DiffusersSuccess8,940.3 s94.5 GB peakZenn: Mochi 1 on M4 Max 64GBPublished 2026-09-02 · Grade A

Which Mac software routes are listed?

  • ComfyUI

    Official

    Formats: safetensors, gguf. Memory policy: component stages stay resident.

    On Apple Silicon ComfyUI runs on PyTorch MPS, not MLX. GGUF files need the separate city96/ComfyUI-GGUF custom node. Resident size is the sum of the weights, an upper bound: ComfyUI can unload models between stages and publishes no per-model Mac peak.

    Sources: ComfyUI README · ComfyUI system requirements · ComfyUI CLI arguments · ComfyUI-GGUF README

  • Diffusers

    Official

    Formats: safetensors, gguf. Memory policy: component stages stay resident.

    Runs on the PyTorch mps device on Mac and CUDA on NVIDIA. GGUF loads through from_single_file on the model class, not the pipeline. Resident size is the sum of the weights, an upper bound.

    Sources: Diffusers MPS guide · Diffusers GGUF guide

What does the official configuration say?

"Official Diffusers BF16 example: 22GB VRAM (CUDA). Single-GPU operation of the genmo harness needs about 60GB VRAM. No Mac statement."

Source: Mochi 1 model card · Read 2026-10-24.

What known issues should you check?

  • Known issue 1

    The only complete Mac measurement ran with 19.8 GB of swap on a 64 GB machine.

    Source: Zenn: Mochi 1 on M4 Max 64GB · Read 2026-10-24

  • Known issue 2

    The official card calls Mochi 1 a living and evolving checkpoint under the research preview, with 480p output.

    Source: Mochi 1 model card · Read 2026-10-24

Frequently asked questions

Can a 64 GB Mac run Mochi 1?

It completed, but slowly. Diffusers MochiPipeline with the bf16 variant on an M4 Max with 64 GB took 8940.3 s for 848x480, 84 frames and 64 steps. Peak unified memory reached 94.5 GB, and 19.8 GB of swap was reported.

What does the official card say about memory?

The card says single-GPU operation needs about 60GB VRAM, the full-precision Diffusers example needs 42GB and the BF16 example needs 22GB. These are CUDA figures for different configurations, not a Mac minimum.

Sources