Is gemma-4-12B-it or MiniCPM-V-4 smaller?

MiniCPM-V-4 is smaller at Q4_K_M: 2.04 GiB against 6.63 GiB.

Which uses less memory at long context?

MiniCPM-V-4 — its KV cache at 32K context is 2.5x smaller, because of how its attention layers cache.

Model comparison

gemma-4-12B-it vs MiniCPM-V-4

At Q4_K_M, MiniCPM-V-4 is the smaller download — 2,189,861,216 bytes against 7,121,861,440. At long context the gap widens: MiniCPM-V-4's KV cache at 32K is 2.5× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	gemma-4-12B-it	MiniCPM-V-4
Parameters	12.0B	4.1B
Architecture	gemma4	llama
Layers	48	32
Native context	262,144	32,768
Mixture of experts	no	no
Quantizations published	59	10
Smallest quantization	3.92 GiB	1.94 GiB
Q4_K_M	6.63 GiB	2.04 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	gemma-4-12B-it	MiniCPM-V-4	Ratio
4,096	0.72 GiB	0.13 GiB	5.75×
8,192	0.97 GiB	0.25 GiB	3.88×
16,384	1.47 GiB	0.50 GiB	2.94×
32,768	2.47 GiB	1.00 GiB	2.47×
65,536	4.47 GiB	2.00 GiB	2.23×
131,072	8.47 GiB	4.00 GiB	2.12×