Is gemma-4-E4B-it-assistant or gemma-3-270m-it smaller?

gemma-4-E4B-it-assistant is smaller at Q4_K_M: 0.07 GiB against 0.24 GiB.

Which uses less memory at long context?

gemma-4-E4B-it-assistant — its KV cache at 32K context is 1.6x smaller, because of how its attention layers cache.

Model comparison

gemma-4-E4B-it-assistant vs gemma-3-270m-it

At Q4_K_M, gemma-4-E4B-it-assistant is the smaller download — 77,231,200 bytes against 253,115,168. At long context the gap widens: gemma-4-E4B-it-assistant's KV cache at 32K is 1.6× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	gemma-4-E4B-it-assistant	gemma-3-270m-it
Parameters	79M	268M
Architecture	gemma4-assistant	gemma3
Layers	4	18
Native context	131,072	32,768
Mixture of experts	no	no
Quantizations published	9	39
Smallest quantization	0.07 GiB	0.17 GiB
Q4_K_M	0.07 GiB	0.24 GiB
Licence	apache-2.0	gemma

KV cache by context

the term that decides long-context viability

Context	gemma-4-E4B-it-assistant	gemma-3-270m-it	Ratio
4,096	0.01 GiB	0.03 GiB	1.93×
8,192	0.02 GiB	0.04 GiB	1.77×
16,384	0.04 GiB	0.06 GiB	1.66×
32,768	0.07 GiB	0.11 GiB	1.59×
65,536	0.13 GiB	0.20 GiB	1.54×
131,072	0.26 GiB	0.39 GiB	1.52×