Is DeepSeek-R1-Distill-Qwen-7B-Uncensored or gemma-4-E4B-it smaller?

DeepSeek-R1-Distill-Qwen-7B-Uncensored is smaller at Q4_K_M: 4.36 GiB against 4.64 GiB.

Which uses less memory at long context?

gemma-4-E4B-it — its KV cache at 32K context is 3.5x smaller, because of how its attention layers cache.

Model comparison

DeepSeek-R1-Distill-Qwen-7B-Uncensored vs gemma-4-E4B-it

At Q4_K_M, DeepSeek-R1-Distill-Qwen-7B-Uncensored is the smaller download — 4,681,087,776 bytes against 4,977,171,584. At long context the gap widens: gemma-4-E4B-it's KV cache at 32K is 3.5× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	DeepSeek-R1-Distill-Qwen-7B-Uncensored	gemma-4-E4B-it
Parameters	7.6B	8.0B
Architecture	qwen2	gemma4
Layers	28	42
Native context	131,072	131,072
Mixture of experts	no	no
Quantizations published	36	36
Smallest quantization	1.77 GiB	3.30 GiB
Q4_K_M	4.36 GiB	4.64 GiB
Licence	mit	apache-2.0

KV cache by context

the term that decides long-context viability

Context	DeepSeek-R1-Distill-Qwen-7B-Uncensored	gemma-4-E4B-it	Ratio
4,096	0.22 GiB	0.12 GiB	1.78×
8,192	0.44 GiB	0.18 GiB	2.46×
16,384	0.88 GiB	0.29 GiB	3.05×
32,768	1.75 GiB	0.51 GiB	3.46×
65,536	3.50 GiB	0.94 GiB	3.71×
131,072	7.00 GiB	1.82 GiB	3.85×