How much VRAM does Kimi K2.5 need?

Kimi K2.5 (1000B parameters) requires approximately 635.4 GB of memory with Q4_K_M quantization.

What is the best quantization for Kimi K2.5?

The recommended quantization for Kimi K2.5 is Q4_K_M, which balances quality and memory efficiency.

Can it run?

No, Kimi K2.5 requires more memory than NVIDIA B200 180GB provides.

FWon't run

Too heavy

Using Q4_K_M in vLLM

Capabilities:

Fit status

Too heavy

Decode

29.0 tok/s

TTFT

6676 ms

Safe context

Memory

635.4 GB / 180.0 GB

Weights610.0 GB

KV Cache5.0 GB

Runtime2.4 GB

Headroom18.0 GB

Workload	Grade	Fit	Decode	TTFT	Context
Agentic Coding	F	Too heavy	33.8 tok/s	8323 ms	9K
Chat	F	Too heavy	33.8 tok/s	3121 ms	4K
Coding	F	Too heavy	29.0 tok/s	6676 ms	5K
RAG	F	Too heavy	33.8 tok/s	10404 ms	9K
Reasoning	F	Too heavy	33.8 tok/s	6762 ms	5K

How Kimi K2.5 (1000B params) fits at each quantization level on NVIDIA B200 180GB (180.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	390.0 GB	Low	F0
Q3_K_S	3	490.0 GB	Low	F0
NVFP4	4