Is MATE-3B or Qwen3-4B smaller?

MATE-3B is smaller at Q4_K_M: 1.80 GiB against 2.33 GiB.

Which uses less memory at long context?

MATE-3B — its KV cache at 32K context is 4.0x smaller, because of how its attention layers cache.

Model comparison

MATE-3B vs Qwen3-4B

At Q4_K_M, MATE-3B is the smaller download — 1,929,903,072 bytes against 2,497,280,256. At long context the gap widens: MATE-3B's KV cache at 32K is 4.0× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	MATE-3B	Qwen3-4B
Parameters	3.1B	4.0B
Architecture	qwen2	qwen3
Layers	36	36
Native context	32,768	40,960
Mixture of experts	no	no
Quantizations published	36	29
Smallest quantization	0.74 GiB	1.01 GiB
Q4_K_M	1.80 GiB	2.33 GiB
Licence	cc-by-nc-nd-4.0	—

KV cache by context

the term that decides long-context viability

Context	MATE-3B	Qwen3-4B	Ratio
4,096	0.14 GiB	0.56 GiB	4.00×
8,192	0.28 GiB	1.13 GiB	4.00×
16,384	0.56 GiB	2.25 GiB	4.00×
32,768	1.13 GiB	4.50 GiB	4.00×
65,536	2.25 GiB	9.00 GiB	4.00×
131,072	4.50 GiB	18.00 GiB	4.00×