Huihui3.5 67B A3B Abliterated needs about 41.9 GB of memory at Q4_K_M (a 40.8 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 11 quants, from 24.7 GB (Q2_K) to 71.2 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 2.7 GB.
Huihui3.5 67B A3B Abliterated – 67B parameter Qwen35MOE model
Huihui3.5 67B A3B Abliterated is a 67B parameter Qwen35MOE model with 3B active experts. It supports vision and is licensed under apache-2.0. The model was trained with a context of 262144 tokens.
Good at
- Vision tasks
- Large context reasoning
- Mixture of experts efficiency
Limits
- No text-only usage limit
- Requires high VRAM for full precision
- License allows redistribution
Best for
- Image captioning
- Long-form document analysis
- Multimodal chat
Who it’s for. Developers needing a large multimodal model with vision support and a permissive license. Researchers and hobbyists who can manage the memory requirements for a 67B parameter model.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 24.7 GB | 25.9 GB | Too big | Get file → | |
| Q3_K_S | 29.1 GB | 30.3 GB | Too big | Get file → | |
| Q3_K_M | 32.2 GB | 33.4 GB | Too big | Get file → | |
| Q3_K_L | 34.8 GB | 36.0 GB | Too big | Get file → | |
| IQ4_XS | 36.3 GB | 37.5 GB | Too big | Get file → | |
| Q4_K_S | 38.3 GB | 39.4 GB | Too big | Get file → | |
| Q4_K_M | 40.8 GB | 41.9 GB | Too big | Get file → | |
| Q5_K_S | 46.2 GB | 47.4 GB | Too big | Get file → | |
| Q5_K_M | 47.7 GB | 48.8 GB | Too big | Get file → | |
| Q6_K | 55.0 GB | 56.2 GB | Too big | Get file → | |
| Q8_0 | 71.2 GB | 72.4 GB | Too big | Get file → |
For 8 GB VRAM, use Q2_K (24.68 GB file, 25.9 GB memory at 8 k context). For 16 GB VRAM, Q3_K_S (29.11 GB file, 30.3 GB memory) or Q4_K_S (38.26 GB file, 39.4 GB memory). For 32 GB VRAM, Q5_K_M (47.67 GB file, 48.8 GB memory) or Q6_K (55.02 GB file, 56.2 GB memory).
What GPU do you need for Huihui3.5 67B A3B Abliterated?
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: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: no quant fits at an 8K context.
- 32 GB VRAM: the best quant that fits is Q3_K_S (29.1 GB file, about 30.3 GB in use).
- 48 GB VRAM: the best quant that fits is Q5_K_S (46.2 GB file, about 47.4 GB in use).
Which Huihui3.5 67B A3B Abliterated quant should you choose?
Q4_K_M is the usual starting point for Huihui3.5 67B A3B Abliterated: it balances file size and quality well. Here it is a 40.8 GB file.
The Q8_0 file is 71.2 GB, about 1.7 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is Q2_K at 24.7 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 Huihui3.5 67B A3B Abliterated
How much VRAM does Huihui3.5 67B A3B Abliterated need?
About 41.9 GB at Q4_K_M with an 8K context window. That is the 40.8 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 Huihui3.5 67B A3B Abliterated on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 25.9 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Huihui3.5 67B A3B Abliterated GGUF?
It depends on the quant: Q2_K is 24.7 GB, Q3_K_S is 29.1 GB, Q3_K_M is 32.2 GB, Q3_K_L is 34.8 GB, IQ4_XS is 36.3 GB, Q4_K_S is 38.3 GB and 5 more.
What context length does Huihui3.5 67B A3B Abliterated support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.7 GB at 8K and 2.7 GB at 32K.
Where can I download Huihui3.5 67B A3B Abliterated GGUF?
The files are on Hugging Face in the mradermacher/Huihui3.5-67B-A3B-abliterated-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 Huihui3.5 67B A3B Abliterated 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 Huihui3.5 67B A3B Abliterated on 8 GB VRAM?
Only the Q2_K quant (24.68 GB file) can be loaded, but it still requires about 25.9 GB memory at 8 k context, so 8 GB VRAM is insufficient.
Does this model support vision?
Yes, the model has detected vision abilities.
What license applies to this model?
The model is licensed under apache-2.0.