dillfrescott_sonya-medium-x8-moe needs about 41.7 GB of memory at Q4_K_M (a 39.6 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 6 quants, from 23.4 GB (Q2_K) to 48.2 GB (Q5_K_M).
Sonya Medium x8 MoE – 69.9B parameter Llama model in GGUF format
Sonya Medium x8 MoE is a 69.9B parameter Llama architecture model quantised for GGUF. It supports a 8192 token context window and is available in several quantisation levels. The model is released under the wtfpl license.
Good at
- Large context handling
- Mixture of experts architecture
- High parameter count
Limits
- Requires large VRAM for higher quants
- No vision capabilities
- License is wtfpl
Best for
- Chat and support applications
- Local LLM deployment
- Text generation with extended context
Who it’s for. Developers and researchers needing a large Llama model for local inference. The model is suitable for users with GPU resources to run higher‑quant versions.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 23.4 GB | 25.5 GB | Too big | Get file → | |
| Q3_K_M | 30.5 GB | 32.6 GB | Too big | Get file → | |
| Q4_0 | 39.6 GB | 41.7 GB | Too big | Get file → | |
| Q4_K_M | 39.6 GB | 41.7 GB | Too big | Get file → | |
| Q5_0 | 48.2 GB | 50.4 GB | Too big | Get file → | |
| Q5_K_M | 48.2 GB | 50.4 GB | Too big | Get file → |
For 8 GB VRAM, use Q2_K (23.39 GB file, ~25.5 GB memory). For 16 GB VRAM, Q3_K_M (30.46 GB file, ~32.6 GB memory). For 32 GB VRAM, Q5_K_M (48.25 GB file, ~50.4 GB memory). Estimates are approximate.
What GPU do you need for dillfrescott_sonya-medium-x8-moe?
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 Q2_K (23.4 GB file, about 25.5 GB in use).
- 48 GB VRAM: the best quant that fits is Q4_0 (39.6 GB file, about 41.7 GB in use).
Which dillfrescott_sonya-medium-x8-moe quant should you choose?
Q4_K_M is the usual starting point for dillfrescott_sonya-medium-x8-moe: it balances file size and quality well. Here it is a 39.6 GB file.
If memory is tight, the smallest file is Q2_K at 23.4 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 dillfrescott_sonya-medium-x8-moe
How much VRAM does dillfrescott_sonya-medium-x8-moe need?
About 41.7 GB at Q4_K_M with an 8K context window. That is the 39.6 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 dillfrescott_sonya-medium-x8-moe on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 25.5 GB, so you would need a larger GPU or CPU offloading.
What is the file size of dillfrescott_sonya-medium-x8-moe GGUF?
It depends on the quant: Q2_K is 23.4 GB, Q3_K_M is 30.5 GB, Q4_0 is 39.6 GB, Q4_K_M is 39.6 GB, Q5_0 is 48.2 GB, Q5_K_M is 48.2 GB.
What context length does dillfrescott_sonya-medium-x8-moe support?
The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 1.6 GB at 8K and 1.6 GB at 8K.
Where can I download dillfrescott_sonya-medium-x8-moe GGUF?
The files are on Hugging Face in the TheBloke/sonya-medium-x8-MoE-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 dillfrescott_sonya-medium-x8-moe use?
Hugging Face lists the license as wtfpl. 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 this model on 8 GB VRAM?
Yes, the Q2_K quantisation requires about 25.5 GB memory, which is close to 8 GB VRAM when using CPU or low‑precision settings.
Does this model support image input?
No, the model has no vision capabilities and only processes text.
What license does it use?
The model is released under the wtfpl license, allowing free use and modification.