Minimax H3 pruned needs about 16.9 GB of memory at Q4_K_S (a 14.3 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 20 quants, from 6.3 GB (IQ1_M) to 50.3 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 8.6 GB.
MiniMax H3 pruned – 20.1B parameter Qwen3VL model in GGUF format
MiniMax H3 pruned is a 20.1B parameter Qwen3VL model repackaged for GGUF. It supports text, image, and video inference. The model is licensed under apache-2.0 and is available in multiple quantizations.
Good at
- Text, image, and video inference
- Supports large context up to 262144 tokens
- Compatible with llama.cpp and koboldcpp
Limits
- No vision support in text-only contexts
- Requires 8k context for memory estimates
- Large VRAM needed for higher quants
Best for
- Text-to-video generation
- Image-to-video conversion
- Video editing workflows
Who it’s for. Developers and researchers building multimodal applications who need a large Qwen3VL model in GGUF format. The model is suitable for those who can manage the VRAM requirements of higher quantizations.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_M | 6.3 GB | 8.9 GB | Too big | Get file → | |
| Q2_K | 6.7 GB | 9.3 GB | Too big | Get file → | |
| Q3_K | 8.7 GB | 11.4 GB | Too big | Get file → | |
| Q2_K_XL | 9.6 GB | 12.2 GB | Too big | Get file → | |
| Q4_0 | 11.4 GB | 14.0 GB | Too big | Get file → | |
| Q4_K | 11.4 GB | 14.0 GB | Too big | Get file → | |
| Q3_K_XL | 12.4 GB | 15.0 GB | Too big | Get file → | |
| Q5_K | 13.9 GB | 16.5 GB | Too big | Get file → | |
| IQ4_NL | 14.2 GB | 16.8 GB | Too big | Get file → | |
| Q4_K_S | 14.3 GB | 16.9 GB | Too big | Get file → | |
| Q4_0 | 14.4 GB | 17.0 GB | Too big | Get file → | |
| Q4_K_XL | 15.1 GB | 17.8 GB | Too big | Get file → | |
| Q6_K | 16.6 GB | 19.2 GB | Too big | Get file → | |
| Q5_K_XL | 17.7 GB | 20.3 GB | Too big | Get file → | |
| Q5_K_XL | 18.3 GB | 20.9 GB | Too big | Get file → | |
| Q8_0 | 21.4 GB | 24.1 GB | Too big | Get file → | |
| Q6_K_XL | 22.2 GB | 24.8 GB | Too big | Get file → | |
| Q8_0 | 26.7 GB | 29.4 GB | Too big | Get file → | |
| BF16 | 40.2 GB | 42.9 GB | Too big | Get file → | |
| BF16 | 50.3 GB | 53.0 GB | Too big | Get file → |
For 8 GB VRAM, use IQ1_M (file size 6.26 GB, memory 8.9 GB at 8k context) or Q2_K (6.68 GB, 9.3 GB). For 16 GB VRAM, Q4_K (11.38 GB, 14 GB) or Q5_K (13.89 GB, 16.5 GB) provide a good balance of quality and memory. For 24 GB VRAM, Q6_K (16.55 GB, 19.2 GB) or Q6_K_XL (22.17 GB, 24.8 GB) are suitable.
What GPU do you need for Minimax H3 pruned?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) 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: the best quant that fits is Q3_K (8.7 GB file, about 11.4 GB in use).
- 16 GB VRAM: the best quant that fits is Q3_K_XL (12.4 GB file, about 15.0 GB in use).
- 24 GB VRAM: the best quant that fits is Q5_K_XL (18.3 GB file, about 20.9 GB in use).
- 32 GB VRAM: the best quant that fits is Q8_0 (26.7 GB file, about 29.4 GB in use).
- 48 GB VRAM: the best quant that fits is BF16 (40.2 GB file, about 42.9 GB in use).
Which Minimax H3 pruned quant should you choose?
Q4_K_S (14.3 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 IQ1_M at 6.3 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Minimax H3 pruned
How much VRAM does Minimax H3 pruned need?
About 16.9 GB at Q4_K_S with an 8K context window. That is the 14.3 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 pruned on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 8.9 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Minimax H3 pruned GGUF?
It depends on the quant: IQ1_M is 6.3 GB, Q2_K is 6.7 GB, Q3_K is 8.7 GB, Q2_K_XL is 9.6 GB, Q4_0 is 11.4 GB, Q4_K is 11.4 GB and 14 more.
What context length does Minimax H3 pruned support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 2.1 GB at 8K and 8.6 GB at 32K.
Where can I download Minimax H3 pruned GGUF?
The files are on Hugging Face in the PurpleBlazey/Minimax-H3-pruned-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 pruned use?
Hugging Face lists the license as apache-2.0. 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 pruned on 8 GB VRAM?
Yes, use IQ1_M or Q2_K quantizations. They require about 8.9 to 9.3 GB of VRAM at 8k context.
What is the best quant for video generation?
Q5_K offers a quality sweet spot while keeping VRAM usage manageable for video resolution and length.
Is the model licensed for commercial use?
The model is licensed under apache-2.0, which allows commercial use with attribution.