Is Huihui-gemma-4-E2B-it-qat-q4_0-unquantized-abliterated or Qwen3-4B smaller?

Qwen3-4B is smaller at Q4_K_M: 2.33 GiB against 3.18 GiB.

Which uses less memory at long context?

Huihui-gemma-4-E2B-it-qat-q4_0-unquantized-abliterated — its KV cache at 32K context is 18.3x smaller, because of how its attention layers cache.

Model comparison

Huihui-gemma-4-E2B-it-qat-q4_0-unquantized-abliterated vs Qwen3-4B

At Q4_K_M, Qwen3-4B is the smaller download — 2,497,280,256 bytes against 3,416,119,296. At long context the gap widens: Huihui-gemma-4-E2B-it-qat-q4_0-unquantized-abliterated's KV cache at 32K is 18.3× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Huihui-gemma-4-E2B-it-qat-q4_0-unquantized-abliterated	Qwen3-4B
Parameters	5.1B	4.0B
Architecture	gemma4	qwen3
Layers	35	36
Native context	131,072	40,960
Mixture of experts	no	no
Quantizations published	28	29
Smallest quantization	2.16 GiB	1.01 GiB
Q4_K_M	3.18 GiB	2.33 GiB
Licence	apache-2.0	—

KV cache by context

the term that decides long-context viability

Context	Huihui-gemma-4-E2B-it-qat-q4_0-unquantized-abliterated	Qwen3-4B	Ratio
4,096	0.05 GiB	0.56 GiB	10.29×
8,192	0.08 GiB	1.13 GiB	13.71×
16,384	0.14 GiB	2.25 GiB	16.46×
32,768	0.25 GiB	4.50 GiB	18.29×
65,536	0.46 GiB	9.00 GiB	19.36×
131,072	0.90 GiB	18.00 GiB	19.95×