Kimi-K2-Thinking — which GPUs fit

Moonshot · 1058.1B (active 32B) · MoE · ctx 256K · Modified MIT · HuggingFace

Kimi-K2-Thinking (Moonshot) is a 1058.1B parameters MoE (32B active per token) open-weight model tracked on AI Hashrate. At 8K context we estimate roughly 589.36 GB VRAM for Q4 (weights 581.96 GB + KV 6.4 GB + 1 GB overhead) and 2123.6 GB for FP16. About 0 GPUs/accelerators in our catalog fully fit this model at Q4 under that context assumption. Fastest Q4 configs: B200 192GB SXM ≈ 48.2 tok/s (estimated); MI300X 192GB ≈ 32.0 tok/s (estimated); Mac Studio M2 Ultra 192GB ≈ 4.8 tok/s (estimated). Use the table below for VRAM fit and tok/s per dollar (list MSRP). Estimates follow memory-bandwidth math; measured rows override when present. See methodology for details. Methodology.

3 hardware configs — sorted by tok/s. Default fit context 8K.

Hardware	VRAM	Quant	tok/s	Fits?	tok/s/$
B200 192GB SXM	192	Q4	48.2 est.	No	0.001
MI300X 192GB	192	Q4	32.0 est.	No	0.002
Mac Studio M2 Ultra 192GB	192	Q4	4.8 est.	No	0.001