Hy3 needs about 185.5 GB of memory at Q4_K_M (a 182.3 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 8 quants, from 89.4 GB (IQ1_M) to 184.7 GB (Q4_K_M). Longer context needs more: at 32K tokens the KV cache alone adds about 10.7 GB.
Hy3: 300B-parameter Mixture-of-Experts model with 8‑k context
Hy3 is a 300‑billion‑parameter Mixture‑of‑Experts model based on the hy_v3 architecture. It supports a 262,144‑token context window and is released under the Apache‑2.0 license. The model is available in several GGUF quantizations for llama.cpp.
Good at
- Large‑context reasoning
- Agentic task performance
- Productivity coding
Limits
- Requires 8‑k context for full capability
- Large memory footprint for high‑quality quants
- No vision support
Best for
- Long‑form document analysis
- Complex coding assistance
- Agent‑based workflow automation
Who it’s for. Developers and researchers needing a high‑capacity, large‑context LLM for text‑only workloads. The model is suitable for production environments that can accommodate its memory requirements.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_M | 89.4 GB | 92.6 GB | Too big | Get file → | |
| IQ1_M | 91.8 GB | 94.9 GB | Too big | Get file → | |
| Q2_K_XL | 100.1 GB | 103.3 GB | Too big | Get file → | |
| Q2_K_XL | 101.6 GB | 104.8 GB | Too big | Get file → | |
| Q3_K_M | 134.6 GB | 137.8 GB | Too big | Get file → | |
| Q3_K_M | 136.5 GB | 139.6 GB | Too big | Get file → | |
| Q4_K_M | 182.3 GB | 185.5 GB | Too big | Get file → | |
| Q4_K_M | 184.7 GB | 187.8 GB | Too big | Get file → |
For 8 GB VRAM use the IQ1_M quant (~89 GB file, ~93 GB memory at 8 k context). If 16 GB VRAM is available, Q2_K_XL (~100 GB file, ~104 GB memory) is a good balance. For larger GPUs (24 GB+), Q4_K_M (~182 GB file, ~186 GB memory) offers higher quality.
What GPU do you need for Hy3?
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: no quant fits at an 8K context.
- 48 GB VRAM: no quant fits at an 8K context.
Which Hy3 quant should you choose?
Q4_K_M is the usual starting point for Hy3: it balances file size and quality well. Here it is a 182.3 GB file.
If memory is tight, the smallest file is IQ1_M at 89.4 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 Hy3
How much VRAM does Hy3 need?
About 185.5 GB at Q4_K_M with an 8K context window. That is the 182.3 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 Hy3 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 92.6 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Hy3 GGUF?
It depends on the quant: IQ1_M is 89.4 GB, IQ1_M is 91.8 GB, Q2_K_XL is 100.1 GB, Q2_K_XL is 101.6 GB, Q3_K_M is 134.6 GB, Q3_K_M is 136.5 GB and 2 more.
What context length does Hy3 support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 2.7 GB at 8K and 10.7 GB at 32K.
Where can I download Hy3 GGUF?
The files are on Hugging Face in the AngelSlim/Hy3-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 Hy3 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 Hy3 on 8 GB VRAM?
Only the IQ1_M quant can be run on 8 GB VRAM, but it requires a very small context and no MTP. Larger quants need more VRAM.
Does Hy3 support vision tasks?
No, Hy3 is a text‑only model and does not include vision capabilities.