Mixtral_7Bx4_MOE_24B needs about 16.2 GB of memory at Q4_K_M (a 14.6 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 21 quants, from 8.8 GB (Q2_K) to 19.8 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 4.3 GB.
Mixtral_7Bx4_MOE_24B is a 24.2B-parameter Mistral model in GGUF format, quantized by RichardErkhov. Its trained context length is 32K tokens. The repository has 21 quant files, from 8.8 GB (Q2_K) to 19.8 GB (Q6_K).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 8.8 GB | 10.4 GB | Too big | Get file → | |
| IQ3_XS | 9.9 GB | 11.5 GB | Too big | Get file → | |
| Q3_K_S | 10.4 GB | 12.0 GB | Too big | Get file → | |
| IQ3_S | 10.5 GB | 12.0 GB | Too big | Get file → | |
| IQ3_M | 10.6 GB | 12.2 GB | Too big | Get file → | |
| Q3_K | 11.6 GB | 13.2 GB | Too big | Get file → | |
| Q3_K_M | 11.6 GB | 13.2 GB | Too big | Get file → | |
| Q3_K_L | 12.5 GB | 14.1 GB | Too big | Get file → | |
| IQ4_XS | 13.0 GB | 14.6 GB | Too big | Get file → | |
| Q4_0 | 13.6 GB | 15.2 GB | Too big | Get file → | |
| Q4_K_S | 13.7 GB | 15.3 GB | Too big | Get file → | |
| IQ4_NL | 13.8 GB | 15.3 GB | Too big | Get file → | |
| Q4_K | 14.6 GB | 16.2 GB | Too big | Get file → | |
| Q4_K_M | 14.6 GB | 16.2 GB | Too big | Get file → | |
| Q4_1 | 15.1 GB | 16.7 GB | Too big | Get file → | |
| Q5_0 | 16.6 GB | 18.2 GB | Too big | Get file → | |
| Q5_K_S | 16.6 GB | 18.2 GB | Too big | Get file → | |
| Q5_K | 17.1 GB | 18.7 GB | Too big | Get file → | |
| Q5_K_M | 17.1 GB | 18.7 GB | Too big | Get file → | |
| Q5_1 | 18.1 GB | 19.7 GB | Too big | Get file → | |
| Q6_K | 19.8 GB | 21.4 GB | Too big | Get file → |
What GPU do you need for Mixtral_7Bx4_MOE_24B?
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 IQ3_S (10.5 GB file, about 12.0 GB in use).
- 16 GB VRAM: the best quant that fits is IQ4_NL (13.8 GB file, about 15.3 GB in use).
- 24 GB VRAM: the best quant that fits is Q6_K (19.8 GB file, about 21.4 GB in use).
Which Mixtral_7Bx4_MOE_24B quant should you choose?
Q4_K_M is the usual starting point for Mixtral_7Bx4_MOE_24B: it balances file size and quality well. Here it is a 14.6 GB file.
If memory is tight, the smallest file is Q2_K at 8.8 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
Frequently asked questions about Mixtral_7Bx4_MOE_24B
How much VRAM does Mixtral_7Bx4_MOE_24B need?
About 16.2 GB at Q4_K_M with an 8K context window. That is the 14.6 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 Mixtral_7Bx4_MOE_24B on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 10.4 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Mixtral_7Bx4_MOE_24B GGUF?
It depends on the quant: Q2_K is 8.8 GB, IQ3_XS is 9.9 GB, Q3_K_S is 10.4 GB, IQ3_S is 10.5 GB, IQ3_M is 10.6 GB, Q3_K is 11.6 GB and 15 more.
What context length does Mixtral_7Bx4_MOE_24B support?
The GGUF header lists a trained context length of 32,768 tokens (32K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.3 GB at 32K.
Where can I download Mixtral_7Bx4_MOE_24B GGUF?
The files are on Hugging Face in the RichardErkhov/cloudyu_-_Mixtral_7Bx4_MOE_24B-gguf repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.