MiniMax-H3 fl2va needs about 22.7 GB of memory at Q4_0 (a 19.9 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 4 quants, from 19.9 GB (Q4_0) to 25.9 GB (Q5_1).
MiniMax-H3 fl2va: 33.1B multimodal LLM for text, video, and audio generation
MiniMax-H3 fl2va is a 33.1B parameter model that supports text, video, and audio generation. It is quantized for use with llama.cpp and ComfyUI. The model is available in Q4_0 and Q5_1 quantizations.
Good at
- Text to video and audio generation
- Reference-conditioned generation
- Multimodal context understanding
Limits
- Requires large VRAM for Q5_1 (≈29.5 GB at 8k context)
- No vision support in the GGUF file
- License is other, not specified
Best for
- Generating short video clips with audio
- Creating reference-conditioned media
- Running in ComfyUI with GGUF support
Who it’s for. Developers and researchers needing a multimodal generative model for text, video, and audio. The model is suitable for those who can provide the necessary GPU memory and have ComfyUI installed.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_0 | 19.9 GB | 22.7 GB | Too big | Get file → | |
| Q4_0 | 19.9 GB | 22.7 GB | Too big | Get file → | |
| Q5_1 | 25.9 GB | 29.5 GB | Too big | Get file → | |
| Q5_1 | 25.9 GB | 29.5 GB | Too big | Get file → |
For 8 GB VRAM use the Q4_0 quant (file size 19.86 GB, estimated 22.7 GB memory at 8k context). For 16 GB VRAM use Q5_1 quant (file size 25.92 GB, estimated 29.5 GB memory at 8k context).
What GPU do you need for MiniMax-H3 fl2va?
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: no quant fits at an 8K context.
- 24 GB VRAM: the best quant that fits is Q4_0 (19.9 GB file, about 22.7 GB in use).
- 32 GB VRAM: the best quant that fits is Q5_1 (25.9 GB file, about 29.5 GB in use).
Which MiniMax-H3 fl2va quant should you choose?
Q4_0 (19.9 GB) is the middle option in this repository. The usual sweet spot between file size and quality, and the quant most people start with.
If memory is tight, the smallest file is Q4_0 at 19.9 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
Source
Frequently asked questions about MiniMax-H3 fl2va
How much VRAM does MiniMax-H3 fl2va need?
About 22.7 GB at Q4_0 with an 8K context window. That is the 19.9 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 fl2va on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 22.7 GB, so you would need a larger GPU or CPU offloading.
What is the file size of MiniMax-H3 fl2va GGUF?
It depends on the quant: Q4_0 is 19.9 GB, Q4_0 is 19.9 GB, Q5_1 is 25.9 GB, Q5_1 is 25.9 GB.
Where can I download MiniMax-H3 fl2va GGUF?
The files are on Hugging Face in the joeygambino/MiniMax-H3-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 fl2va 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 fl2va on 8 GB VRAM?
No. The Q4_0 quant requires approximately 22.7 GB memory at 8k context, which exceeds 8 GB VRAM.
Does the model support vision input?
The GGUF file does not include vision support; it focuses on text, video, and audio generation.
What ComfyUI version is needed?
ComfyUI 0.30.0+ with the ComfyUI-H3-Multishot pack v1.5.2 or newer is required to load the minimax_h3 architecture.