Is DeepSeek-Coder-V2-Lite-Instruct or gemma-4-E4B-it smaller?

gemma-4-E4B-it is smaller at Q4_K_M: 4.64 GiB against 9.65 GiB.

Which uses less memory at long context?

gemma-4-E4B-it — its KV cache at 32K context is 1.9x smaller, because of how its attention layers cache.

Model comparison

DeepSeek-Coder-V2-Lite-Instruct vs gemma-4-E4B-it

At Q4_K_M, gemma-4-E4B-it is the smaller download — 4,977,171,584 bytes against 10,364,416,768. At long context the gap widens: gemma-4-E4B-it's KV cache at 32K is 1.9× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	DeepSeek-Coder-V2-Lite-Instruct	gemma-4-E4B-it
Parameters	15.7B	8.0B
Architecture	deepseek2	gemma4
Layers	27	42
Native context	163,840	131,072
Mixture of experts	yes, 64 experts	no
Quantizations published	45	36
Smallest quantization	4.65 GiB	3.30 GiB
Q4_K_M	9.65 GiB	4.64 GiB
Licence	other	apache-2.0

KV cache by context

the term that decides long-context viability

Context	DeepSeek-Coder-V2-Lite-Instruct	gemma-4-E4B-it	Ratio
4,096	0.12 GiB	0.12 GiB	1.04×
8,192	0.24 GiB	0.18 GiB	1.34×
16,384	0.47 GiB	0.29 GiB	1.65×
32,768	0.95 GiB	0.51 GiB	1.88×
65,536	1.90 GiB	0.94 GiB	2.01×
131,072	3.80 GiB	1.82 GiB	2.09×