Is gemma-7b or gemma-4-E4B-it smaller?

gemma-4-E4B-it is smaller at Q4_K_M: 4.64 GiB against 4.96 GiB.

Which uses less memory at long context?

gemma-4-E4B-it — its KV cache at 32K context is 27.7x smaller, because of how its attention layers cache.

Model comparison

gemma-7b vs gemma-4-E4B-it

At Q4_K_M, gemma-4-E4B-it is the smaller download — 4,977,171,584 bytes against 5,329,758,592. At long context the gap widens: gemma-4-E4B-it's KV cache at 32K is 27.7× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	gemma-7b	gemma-4-E4B-it
Parameters	8.5B	8.0B
Architecture	gemma	gemma4
Layers	28	42
Native context	8,192	131,072
Mixture of experts	no	no
Quantizations published	31	36
Smallest quantization	2.01 GiB	3.30 GiB
Q4_K_M	4.96 GiB	4.64 GiB
Licence	gemma	apache-2.0

KV cache by context

the term that decides long-context viability

Context	gemma-7b	gemma-4-E4B-it	Ratio
4,096	1.75 GiB	0.12 GiB	14.22×
8,192	3.50 GiB	0.18 GiB	19.69×
16,384	7.00 GiB	0.29 GiB	24.38×
32,768	14.00 GiB	0.51 GiB	27.68×
65,536	28.00 GiB	0.94 GiB	29.68×
131,072	56.00 GiB	1.82 GiB	30.80×