MI

MiniMax H3 Turbo

MinimaxGGUF by molbalUpdated 2026-08-13
Download on Hugging Face
Parameters20.1B
Context—
Layers—
Architectureminimax_h3
Smallest quant11.4 GB
Quick answer

MiniMax H3 Turbo needs about 23.1 GB of memory at its main file (a 20.2 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 3 quants, from 11.4 GB (Q4_0) to 21.4 GB (Q8_0).

MiniMax H3 Turbo – 20.1B omni‑modal diffusion model in GGUF format

MiniMax H3 Turbo is a 20.1B parameter diffusion model optimized for 4‑step inference at 768p resolution. It supports text, image, video, and audio generation with native stereo audio. The model is distributed in GGUF quantizations and can be used directly in ComfyUI.

Good at

  • Fast 4‑step generation
  • Omni‑modal text, image, video, audio output
  • High‑resolution 768p results

Limits

  • No vision support in llama.cpp
  • Requires full merged GGUF files
  • Large VRAM needed for higher quantizations

Best for

  • Local diffusion inference
  • ComfyUI workflow integration
  • Rapid prototyping of multimodal content

Who it’s for. Developers and researchers who need a compact, fully merged diffusion model for local inference. The model is suitable for users who prefer GGUF format and ComfyUI integration.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_011.4 GB 13.3 GB
Too big Get file →
—20.2 GB 23.1 GB
Too big Get file →
Q8_021.4 GB 24.5 GB
Too big Get file →

For 8 GB VRAM, use the Q4_0 quantization (11.39 GB file, ~13.3 GB memory at 8k context). If 16 GB VRAM is available, the Q8_0 quantization (21.42 GB file, ~24.5 GB memory) provides higher fidelity. The unknown quant is larger and should be used only with ≥24 GB VRAM.

What GPU do you need for MiniMax H3 Turbo?

These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: the best quant that fits is Q4_0 (11.4 GB file, about 13.3 GB in use).
  • 24 GB VRAM: the best quant that fits is the main file (20.2 GB file, about 23.1 GB in use).
  • 32 GB VRAM: the best quant that fits is Q8_0 (21.4 GB file, about 24.5 GB in use).

Which MiniMax H3 Turbo quant should you choose?

If memory is tight, the smallest file is Q4_0 at 11.4 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureminimax_h3
Parameters20.1B

Source

Quantized bymolbal
FamilyMinimax
Licenseother
Downloads (HF)4,796
Last updated2026-08-13

Frequently asked questions about MiniMax H3 Turbo

How much VRAM does MiniMax H3 Turbo need?

About 23.1 GB at its main file with an 8K context window. That is the 20.2 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run MiniMax H3 Turbo on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 13.3 GB, so you would need a larger GPU or CPU offloading.

What is the file size of MiniMax H3 Turbo GGUF?

It depends on the quant: Q4_0 is 11.4 GB, file is 20.2 GB, Q8_0 is 21.4 GB.

Where can I download MiniMax H3 Turbo GGUF?

The files are on Hugging Face in the molbal/MiniMax-H3-Turbo-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does MiniMax H3 Turbo use?

Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run MiniMax H3 Turbo on 8 GB VRAM?

Yes, using the Q4_0 quantization which requires about 13.3 GB memory at 8k context. It fits within 8 GB VRAM with some memory overhead.

Does the model support vision tasks?

No, the GGUF version does not include vision support in llama.cpp. It is intended for diffusion inference only.

What is the file size of the Q8_0 quant?

The Q8_0 quant file is 21.42 GB, with an estimated memory usage of 24.5 GB at 8k context.

Other GGUFs of MiniMax-H3

Similar models

Browse all →

More from molbal

All by molbal →

More Minimax models

All Minimax →

Other medium models

All medium models →