Kimi-K2-Instruct-0905: hardware requirements

Kimi-K2-Instruct-0905 needs about 632 GB of memory at Q4 quantization — a 621 GB download plus room for context and overhead. That is beyond consumer hardware; use it through an API instead.

Parameters
1026B
Tier
Server only
Licence
See model card
Released
2025-09-05

Will it run on your machine?

Download sizes by quantization

Lower quantization means a smaller file and less memory, at some cost to quality. Q4_K_M is the usual starting point. What is this?

QuantDownloadRAM @ 4KRAM @ 32KSource
Q4_K_Mrecommended621 GB632 GB682 GBmeasured
Q5_K_M729 GB739 GB789 GBmeasured
Q6_K843 GB853 GB903 GBmeasured
Q8_01091 GB1101 GB1152 GBmeasured
BF162053 GB2064 GB2114 GBmeasured

“Measured” sizes are read from published GGUF files (unsloth/Kimi-K2-Instruct-0905-GGUF). “Estimated” sizes are derived from the parameter count and are typically within a few percent. The RAM columns add the context cache and runtime overhead to the download size.