Is Olmo-3-7B-Think or gemma-4-E4B-it smaller?

Olmo-3-7B-Think is smaller at Q4_K_M: 4.16 GiB against 4.64 GiB.

Which uses less memory at long context?

gemma-4-E4B-it — its KV cache at 32K context is 11.2x smaller, because of how its attention layers cache.

Model comparison

Olmo-3-7B-Think vs gemma-4-E4B-it

At Q4_K_M, Olmo-3-7B-Think is the smaller download — 4,471,600,192 bytes against 4,977,171,584. At long context the gap widens: gemma-4-E4B-it's KV cache at 32K is 11.2× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Olmo-3-7B-Think	gemma-4-E4B-it
Parameters	7.3B	8.0B
Architecture	olmo2	gemma4
Layers	32	42
Native context	65,536	131,072
Mixture of experts	no	no
Quantizations published	47	36
Smallest quantization	1.69 GiB	3.30 GiB
Q4_K_M	4.16 GiB	4.64 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	Olmo-3-7B-Think	gemma-4-E4B-it	Ratio
4,096	2.00 GiB	0.12 GiB	16.25×
8,192	2.69 GiB	0.18 GiB	15.12×
16,384	3.69 GiB	0.29 GiB	12.84×
32,768	5.69 GiB	0.51 GiB	11.24×
65,536	9.69 GiB	0.94 GiB	10.27×
131,072	17.69 GiB	1.82 GiB	9.73×