Is gemma-4-E4B-uncensored or gemma-4-12B-it-qat-q4_0-unquantized smaller?

Both publish quantizations; compare the table for the exact sizes.

Which uses less memory at long context?

gemma-4-E4B-uncensored — its KV cache at 32K context is 4.9x smaller, because of how its attention layers cache.

Model comparison

gemma-4-E4B-uncensored vs gemma-4-12B-it-qat-q4_0-unquantized

These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: gemma-4-E4B-uncensored's KV cache at 32K is 4.9× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	gemma-4-E4B-uncensored	gemma-4-12B-it-qat-q4_0-unquantized
Parameters	7.9B	12.0B
Architecture	gemma4	gemma4
Layers	42	48
Native context	131,072	262,144
Mixture of experts	no	no
Quantizations published	36	2
Smallest quantization	3.06 GiB	6.50 GiB
Q4_K_M	4.94 GiB	—
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	gemma-4-E4B-uncensored	gemma-4-12B-it-qat-q4_0-unquantized	Ratio
4,096	0.12 GiB	0.72 GiB	5.84×
8,192	0.18 GiB	0.97 GiB	5.45×
16,384	0.29 GiB	1.47 GiB	5.12×
32,768	0.51 GiB	2.47 GiB	4.88×
65,536	0.94 GiB	4.47 GiB	4.74×
131,072	1.82 GiB	8.47 GiB	4.66×