Hy MT2 30B A3B needs about 19.5 GB of memory at Q4_K_M (a 18.2 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 2 quants, from 18.2 GB (Q4_K_M) to 32.0 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 3.2 GB.
Tencent Hy MT2 30B A3B – 30 Billion‑parameter multilingual translation model
Hy MT2 30B A3B is a 30 Billion‑parameter mixture‑of‑experts model built on the hy_v3 architecture. It supports translation among 33 languages and follows translation instructions in multiple languages. The model was trained on 262,144‑token context windows and released under the Apache‑2.0 license.
Good at
- Fast‑thinking multilingual translation
- Instruction‑following translation
- High‑quality business‑domain translation
Limits
- Requires 33 GB VRAM for Q8_0 at 8 k context
- No vision or audio capabilities
- Large file size limits on low‑storage devices
Best for
- Real‑world business translation
- Domain‑specific translation tasks
- Competition in WMT translation challenges
Who it’s for. Developers and researchers needing high‑quality multilingual translation in 33 languages. Suitable for integration into translation services and competition submissions.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_K_M | 18.2 GB | 19.5 GB | Too big | Get file → | |
| Q8_0 | 32.0 GB | 33.3 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_K_M (18.24 GB file, ~19.5 GB VRAM at 8 k context). For larger GPUs, Q8_0 (31.99 GB file, ~33.3 GB VRAM) provides full precision. Memory estimates are approximate.
What GPU do you need for Hy MT2 30B A3B?
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: the best quant that fits is Q4_K_M (18.2 GB file, about 19.5 GB in use).
- 32 GB VRAM: the best quant that fits is Q4_K_M (18.2 GB file, about 19.5 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_0 (32.0 GB file, about 33.3 GB in use).
Which Hy MT2 30B A3B quant should you choose?
Q4_K_M is the usual starting point for Hy MT2 30B A3B: it balances file size and quality well. Here it is a 18.2 GB file.
The Q8_0 file is 32.0 GB, about 1.8 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.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Hy MT2 30B A3B
How much VRAM does Hy MT2 30B A3B need?
About 19.5 GB at Q4_K_M with an 8K context window. That is the 18.2 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 Hy MT2 30B A3B on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 19.5 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Hy MT2 30B A3B GGUF?
It depends on the quant: Q4_K_M is 18.2 GB, Q8_0 is 32.0 GB.
What context length does Hy MT2 30B A3B support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.8 GB at 8K and 3.2 GB at 32K.
Where can I download Hy MT2 30B A3B GGUF?
The files are on Hugging Face in the tencent/Hy-MT2-30B-A3B-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 Hy MT2 30B A3B 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 Hy MT2 30B A3B on 8 GB VRAM?
Only with the Q4_K_M quantization; it requires about 19.5 GB VRAM at 8 k context, so 8 GB is insufficient.
Does the model support vision or audio input?
No, the model is purely text‑based and does not support vision or audio modalities.
What license applies to this model?
The model is released under the Apache‑2.0 license.