Kimi-K3: hardware requirements

Kimi-K3 needs about 1575 GB of memory at Q4 quantization — a 1561 GB download plus room for context and overhead. That is beyond consumer hardware; use it through an API instead.

Parameters
2780B
Tier
Server only
Licence
modified-mit
Released
2026-06-13

This is a mixture-of-experts model. Only a fraction of its 2780B parameters are used for each token, so it runs faster than its size suggests — but all of it still has to fit in memory. The speed is that of a small model; the memory requirement is that of a large one. More on this →

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
Published1561 GB1575 GB1652 GBmeasured

“Measured” sizes are read from published GGUF files (moonshotai/Kimi-K3). “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.