Kanana GGUF models
152 GGUF models in the Kanana family, quantized by 92 uploaders. Sizes run from 1.29B to 30.7B parameters. The longest trained context is 128K tokens.
About the Kanana GGUF models
Kanana is represented on ggufmodels.com by 152 GGUF models, quantized by 92 uploaders on Hugging Face. The most downloaded are Kanana 1.5 15.7b A3B Instruct, Kakaocorp.Kanana 1.5 8b Instruct 2505 and Kanana 1.5 8b Instruct 2505. Parameter counts run from 1.29B to 30.7B, and the longest trained context window is 128K tokens.
A Q4_K_M file of a Kanana model is between 0.9 to 18.6 GB, which is the quant most people start with. Most files come from mradermacher (30), tensorblock (8) and DevQuasar (7). The quants you will meet most often are Q4_K_M (119), Q5_K_M (52), Q8_0 (51) and Q6_K (49), where the number is how many models offer that quant.
Frequently asked questions
Which Kanana GGUF should I download?
The most downloaded Kanana GGUF here is Kanana 1.5 15.7b A3B Instruct. Pick the size that fits your GPU first, then choose a quant: Q4_K_M is the usual starting point, and Q5_K_M or Q6_K give more quality if you have memory to spare. The VRAM pages below list what fits each GPU.
How much VRAM does a Kanana model need?
Q4_K_M files of Kanana models on this site range from 0.9 to 18.6 GB. Add the KV cache for your context length and about 0.5 GB for the runtime, which usually means 1.5 to 3 GB on top of the file at an 8K context.
Where do Kanana GGUF files come from?
The most common quantizers here are mradermacher, tensorblock and DevQuasar. They convert the original model and publish quantized files on Hugging Face. ggufmodels.com does not host files; every download button opens Hugging Face.