# DiffSynth-Studio offers quantized MiniMax-H3-Image checkpoints for low-VRAM inference

> Original title: Deploy MiniMax-H3-Image on ANY hardware with ultra-low cost! 🤖 https://modelscope.ai/models/DiffSynth-Studio/MiniMax-H3-Image-EntroPack

- Source: ModelScope
- Published: 2026-11-12T10:00:03.000Z
- Bad Influence: https://badinfluence.news/items/cix4bfwwyifpgeltd6iywktkg
- Original: https://x.com/ModelScope2022/status/2120818899109220659

## Summary

DiffSynth-Studio releases EntroPack checkpoints for MiniMax-H3-Image, with a bf16 pruned model for 80GB VRAM, an INT8 model for 24GB, and a 3-bit model for 8GB. ComfyUI workflows are also available in the model repository.

## Post

Deploy MiniMax-H3-Image on ANY hardware with ultra-low cost! 🤖 https://modelscope.ai/models/DiffSynth-Studio/MiniMax-H3-Image-EntroPack

Fast, low-cost inference is now possible with EntroPack. Choose your checkpoint:
🔹 80GB VRAM: bf16_table_pruned.safetensors (Fully lossless pruned model)
🔹 24GB VRAM: dit_8bit_int8.safetensors (Near-lossless INT8 + ConvRot quantization)
🔹 8GB VRAM: dit_extreme_3bpp.safetensors (Extreme 3-bit quantized model)

ComfyUI workflows also available in the model repository. Try it now!

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHW6QkOyb0AAT8Sp.jpg&mode=thumb&exp=1794700800&sig=5ec28e1886d1674b)

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHW6QmGcawAATSBD.jpg&mode=thumb&exp=1794700800&sig=0e34652d8a694bcc)