How much VRAM does Kimi K2.5 need?

Kimi K2.5 (1000B parameters) requires approximately 625.4 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 A800 80GB provides.

FWon't run

Too heavy

Using Q4_K_M in vLLM

Capabilities:

Fit status

Too heavy

Decode

7.0 tok/s

TTFT

27600 ms

Safe context

Memory

625.4 GB / 80.0 GB

Weights610.0 GB

KV Cache5.0 GB

Runtime2.4 GB

Headroom8.0 GB

Workload	Grade	Fit	Decode	TTFT	Context
Agentic Coding	F	Too heavy	7.6 tok/s	37057 ms	4K
Chat	F	Too heavy	7.6 tok/s	13896 ms	4K
Coding	F	Too heavy	7.0 tok/s	27600 ms	4K
RAG	F	Too heavy	7.6 tok/s	46321 ms	4K
Reasoning	F	Too heavy	7.6 tok/s	30109 ms	4K

How Kimi K2.5 (1000B params) fits at each quantization level on NVIDIA A800 80GB (80.0 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