Kanana 2 30b A3B Thinking 2601 needs about 20.0 GB of memory at Q4_K_M (a 18.6 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 5 quants, from 15.9 GB (Q3_K_L) to 32.6 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 3.4 GB.
Kanana 2 30b A3B Thinking 2601 is a 30.7B-parameter Kanana model in GGUF format, quantized by mykor. Its trained context length is 32K tokens. The repository has 5 quant files, from 15.9 GB (Q3_K_L) to 32.6 GB (Q8_0).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q3_K_L | 15.9 GB | 17.3 GB | Too big | Get file → | |
| Q4_K_M | 18.6 GB | 20.0 GB | Too big | Get file → | |
| Q5_K_M | 21.8 GB | 23.2 GB | Too big | Get file → | |
| Q6_K | 25.2 GB | 26.6 GB | Too big | Get file → | |
| Q8_0 | 32.6 GB | 34.0 GB | Too big | Get file → |
What GPU do you need for Kanana 2 30b A3B Thinking 2601?
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: the best quant that fits is Q5_K_M (21.8 GB file, about 23.2 GB in use).
- 32 GB VRAM: the best quant that fits is Q6_K (25.2 GB file, about 26.6 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_0 (32.6 GB file, about 34.0 GB in use).
Which Kanana 2 30b A3B Thinking 2601 quant should you choose?
Q4_K_M is the usual starting point for Kanana 2 30b A3B Thinking 2601: it balances file size and quality well. Here it is a 18.6 GB file.
The Q8_0 file is 32.6 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.
If memory is tight, the smallest file is Q3_K_L at 15.9 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Kanana 2 30b A3B Thinking 2601
How much VRAM does Kanana 2 30b A3B Thinking 2601 need?
About 20.0 GB at Q4_K_M with an 8K context window. That is the 18.6 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 Kanana 2 30b A3B Thinking 2601 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 17.3 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Kanana 2 30b A3B Thinking 2601 GGUF?
It depends on the quant: Q3_K_L is 15.9 GB, Q4_K_M is 18.6 GB, Q5_K_M is 21.8 GB, Q6_K is 25.2 GB, Q8_0 is 32.6 GB.
What context length does Kanana 2 30b A3B Thinking 2601 support?
The GGUF header lists a trained context length of 32,768 tokens (32K). The KV cache grows with the context you use: about 0.9 GB at 8K and 3.4 GB at 32K.
Where can I download Kanana 2 30b A3B Thinking 2601 GGUF?
The files are on Hugging Face in the mykor/kanana-2-30b-a3b-thinking-2601-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.