Is Qwen3.5-4B or gemma-4-E2B-it-qat-q4_0-unquantized-heretic smaller?

Qwen3.5-4B is smaller at Q4_K_M: 2.52 GiB against 3.18 GiB.

Which uses less memory at long context?

gemma-4-E2B-it-qat-q4_0-unquantized-heretic — its KV cache at 32K context is 4.1x smaller, because of how its attention layers cache.

Model comparison

Qwen3.5-4B vs gemma-4-E2B-it-qat-q4_0-unquantized-heretic

At Q4_K_M, Qwen3.5-4B is the smaller download — 2,707,513,696 bytes against 3,416,119,296. At long context the gap widens: gemma-4-E2B-it-qat-q4_0-unquantized-heretic's KV cache at 32K is 4.1× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Qwen3.5-4B	gemma-4-E2B-it-qat-q4_0-unquantized-heretic
Parameters	4.7B	5.1B
Architecture	qwen35	gemma4
Layers	32	35
Native context	262,144	131,072
Mixture of experts	no	no
Quantizations published	57	28
Smallest quantization	1.42 GiB	2.16 GiB
Q4_K_M	2.52 GiB	3.18 GiB
Licence	apache-2.0	gemma

KV cache by context

the term that decides long-context viability

Context	Qwen3.5-4B	gemma-4-E2B-it-qat-q4_0-unquantized-heretic	Ratio
4,096	0.13 GiB	0.05 GiB	2.29×
8,192	0.25 GiB	0.08 GiB	3.05×
16,384	0.50 GiB	0.14 GiB	3.66×
32,768	1.00 GiB	0.25 GiB	4.06×
65,536	2.00 GiB	0.46 GiB	4.30×
131,072	4.00 GiB	0.90 GiB	4.43×