Kimi-K3 needs about 806.6 GB of memory at IQ2_XXS (a 711.1 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 6 quants, from 594.0 GB (IQ1_S) to 1561.2 GB (Q8_K_XL). Longer context needs more: at 32K tokens the KV cache alone adds about 380.3 GB.
Kimi-K3 is a 2780B-parameter Kimi model in GGUF format, quantized by aria220. Its trained context length is 1024K tokens. The repository has 6 quant files, from 594.0 GB (IQ1_S) to 1561.2 GB (Q8_K_XL).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_S | 594.0 GB (14 parts) | 689.6 GB | Too big | Get file → | |
| IQ1_M | 648.9 GB (15 parts) | 744.5 GB | Too big | Get file → | |
| IQ2_XXS | 711.1 GB (16 parts) | 806.6 GB | Too big | Get file → | |
| Q2_K_XL | 861.3 GB (19 parts) | 956.9 GB | Too big | Get file → | |
| Q4_K_XL | 1508.7 GB (32 parts) | 1604.2 GB | Too big | Get file → | |
| Q8_K_XL | 1561.2 GB (34 parts) | 1656.7 GB | Too big | Get file → |
What GPU do you need for Kimi-K3?
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 Kimi-K3 quant should you choose?
IQ2_XXS (711.1 GB) is the middle option in this repository. 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.
If memory is tight, the smallest file is IQ1_S at 594.0 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 Kimi-K3
How much VRAM does Kimi-K3 need?
About 806.6 GB at IQ2_XXS with an 8K context window. That is the 711.1 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 Kimi-K3 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 689.6 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Kimi-K3 GGUF?
It depends on the quant: IQ1_S is 594.0 GB, IQ1_M is 648.9 GB, IQ2_XXS is 711.1 GB, Q2_K_XL is 861.3 GB, Q4_K_XL is 1508.7 GB, Q8_K_XL is 1561.2 GB.
What context length does Kimi-K3 support?
The GGUF header lists a trained context length of 1,048,576 tokens (1024K). The KV cache grows with the context you use: about 95.1 GB at 8K and 380.3 GB at 32K.
Where can I download Kimi-K3 GGUF?
The files are on Hugging Face in the aria220/Kimi-K3-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 Kimi-K3 use?
Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.