MoCha needs about 11.3 GB of memory at Q4_K_M (a 9.7 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 13 quants, from 5.3 GB (Q2_K) to 15.4 GB (Q8_0).
14.3B parameter MoCha model in GGUF format
MoCha is a 14.3B parameter model quantized for llama.cpp. It is available in multiple quantization levels from Q2_K to Q8_0. The model is released under agpl-3.0 license.
Good at
- Large language modeling
- Text generation
- General purpose usage
Limits
- No vision support
- License requires AGPL compliance
- Context length not specified
Best for
- Chatbots with moderate VRAM
- Text summarization
- Code generation
Who it’s for. Developers and researchers who need a large language model that can run on llama.cpp. The model is suitable for those who can comply with the agpl-3.0 license and have sufficient GPU memory.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 5.3 GB | 6.4 GB | Fits | Get file → | |
| Q3_K_S | 6.5 GB | 7.8 GB | Tight | Get file → | |
| Q3_K_M | 7.2 GB | 8.5 GB | Too big | Get file → | |
| Q4_0 | 8.6 GB | 10.1 GB | Too big | Get file → | |
| Q4_K_S | 8.8 GB | 10.3 GB | Too big | Get file → | |
| Q4_1 | 9.3 GB | 10.9 GB | Too big | Get file → | |
| Q4_K_M | 9.7 GB | 11.3 GB | Too big | Get file → | |
| Q5_K_S | 10.1 GB | 11.9 GB | Too big | Get file → | |
| Q5_0 | 10.3 GB | 12.1 GB | Too big | Get file → | |
| Q5_K_M | 10.8 GB | 12.6 GB | Too big | Get file → | |
| Q5_1 | 11.0 GB | 12.8 GB | Too big | Get file → | |
| Q6_K | 12.0 GB | 14.0 GB | Too big | Get file → | |
| Q8_0 | 15.4 GB | 17.8 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_0 (est. 10.1 GB) or Q4_K_S (10.3 GB). For 12 GB VRAM, Q5_0 (12.1 GB) or Q5_K_M (12.6 GB) are suitable. Higher quants like Q6_K or Q8_0 require 14 GB or 17.8 GB respectively.
What GPU do you need for MoCha?
These figures add the file size, an fp16 KV cache at an 8K context window 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: the best quant that fits is Q3_K_S (6.5 GB file, about 7.8 GB in use).
- 12 GB VRAM: the best quant that fits is Q5_K_S (10.1 GB file, about 11.9 GB in use).
- 16 GB VRAM: the best quant that fits is Q6_K (12.0 GB file, about 14.0 GB in use).
- 24 GB VRAM: the best quant that fits is Q8_0 (15.4 GB file, about 17.8 GB in use).
Which MoCha quant should you choose?
Q4_K_M is the usual starting point for MoCha: it balances file size and quality well. Here it is a 9.7 GB file.
The Q8_0 file is 15.4 GB, about 1.6 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 5.3 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 MoCha
How much VRAM does MoCha need?
About 11.3 GB at Q4_K_M with an 8K context window. That is the 9.7 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 MoCha on an 8 GB GPU?
Yes, with Q3_K_S: the 6.5 GB file needs about 7.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of MoCha GGUF?
It depends on the quant: Q2_K is 5.3 GB, Q3_K_S is 6.5 GB, Q3_K_M is 7.2 GB, Q4_0 is 8.6 GB, Q4_K_S is 8.8 GB, Q4_1 is 9.3 GB and 7 more.
Where can I download MoCha GGUF?
The files are on Hugging Face in the vantagewithai/MoCha-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 MoCha use?
Hugging Face lists the license as agpl-3.0. 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 MoCha on 8 GB VRAM?
Yes, using Q4_0 or Q4_K_S quantizations, which estimate 10.1 GB or 10.3 GB memory at 8k context.
Does MoCha support vision tasks?
No, MoCha does not support vision capabilities.