GigaChat3.1 702B A36B needs about 425.5 GB of memory at Q4_K_M (a 423.9 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 4 quants, from 423.9 GB (Q4_K_M) to 1404.3 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 4.6 GB.
GigaChat3.1 702B A36B is a 702B-parameter Gigachat model in GGUF format, quantized by ai-sage. Its trained context length is 256K tokens. The repository has 4 quant files, from 423.9 GB (Q4_K_M) to 1404.3 GB (BF16).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_K_M | 423.9 GB (9 parts) | 425.5 GB | Too big | Get file → | |
| Q6_K | 576.3 GB (13 parts) | 578.0 GB | Too big | Get file → | |
| Q8_0 | 746.3 GB (16 parts) | 747.9 GB | Too big | Get file → | |
| BF16 | 1404.3 GB (31 parts) | 1406.0 GB | Too big | Get file → |
What GPU do you need for GigaChat3.1 702B A36B?
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 GigaChat3.1 702B A36B quant should you choose?
Q4_K_M is the usual starting point for GigaChat3.1 702B A36B: it balances file size and quality well. Here it is a 423.9 GB file.
The Q8_0 file is 746.3 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 GigaChat3.1 702B A36B
How much VRAM does GigaChat3.1 702B A36B need?
About 425.5 GB at Q4_K_M with an 8K context window. That is the 423.9 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 GigaChat3.1 702B A36B on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 425.5 GB, so you would need a larger GPU or CPU offloading.
What is the file size of GigaChat3.1 702B A36B GGUF?
It depends on the quant: Q4_K_M is 423.9 GB, Q6_K is 576.3 GB, Q8_0 is 746.3 GB, BF16 is 1404.3 GB.
What context length does GigaChat3.1 702B A36B support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.6 GB at 32K.
Where can I download GigaChat3.1 702B A36B GGUF?
The files are on Hugging Face in the ai-sage/GigaChat3.1-702B-A36B-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 GigaChat3.1 702B A36B use?
Hugging Face lists the license as mit. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.