Is GigaChat3-10B-A1.8B-base or gemma-4-E4B-it smaller?

gemma-4-E4B-it is smaller at Q4_K_M: 4.64 GiB against 6.03 GiB.

Which uses less memory at long context?

gemma-4-E4B-it — its KV cache at 32K context is 1.8x smaller, because of how its attention layers cache.

Model comparison

GigaChat3-10B-A1.8B-base vs gemma-4-E4B-it

At Q4_K_M, gemma-4-E4B-it is the smaller download — 4,977,171,584 bytes against 6,474,702,976. At long context the gap widens: gemma-4-E4B-it's KV cache at 32K is 1.8× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	GigaChat3-10B-A1.8B-base	gemma-4-E4B-it
Parameters	11.5B	8.0B
Architecture	deepseek2	gemma4
Layers	26	42
Native context	262,144	131,072
Mixture of experts	yes, 64 experts	no
Quantizations published	8	36
Smallest quantization	6.03 GiB	3.30 GiB
Q4_K_M	6.03 GiB	4.64 GiB
Licence	mit	apache-2.0

KV cache by context

the term that decides long-context viability

Context	GigaChat3-10B-A1.8B-base	gemma-4-E4B-it	Ratio
4,096	0.11 GiB	0.12 GiB	1.08×
8,192	0.23 GiB	0.18 GiB	1.29×
16,384	0.46 GiB	0.29 GiB	1.59×
32,768	0.91 GiB	0.51 GiB	1.81×
65,536	1.83 GiB	0.94 GiB	1.94×
131,072	3.66 GiB	1.82 GiB	2.01×