Is MiniCPM-o-4_5 or Qwen3.8-27B smaller?

MiniCPM-o-4_5 is smaller at Q4_K_M: 4.68 GiB against 15.66 GiB.

Which uses less memory at long context?

Qwen3.8-27B — its KV cache at 32K context is 2.2x smaller, because of how its attention layers cache.

Model comparison

MiniCPM-o-4_5 vs Qwen3.8-27B

At Q4_K_M, MiniCPM-o-4_5 is the smaller download — 5,026,714,400 bytes against 16,810,714,336. At long context the gap widens: Qwen3.8-27B's KV cache at 32K is 2.2× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	MiniCPM-o-4_5	Qwen3.8-27B
Parameters	9.4B	27.8B
Architecture	qwen3	qwen35
Layers	36	64
Native context	40,960	262,144
Mixture of experts	no	no
Quantizations published	11	22
Smallest quantization	4.45 GiB	8.39 GiB
Q4_K_M	4.68 GiB	15.66 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	MiniCPM-o-4_5	Qwen3.8-27B	Ratio
4,096	0.56 GiB	0.25 GiB	2.25×
8,192	1.13 GiB	0.50 GiB	2.25×
16,384	2.25 GiB	1.00 GiB	2.25×
32,768	4.50 GiB	2.00 GiB	2.25×
65,536	9.00 GiB	4.00 GiB	2.25×
131,072	18.00 GiB	8.00 GiB	2.25×