Is gemma-2b or Qwen3-4B smaller?

gemma-2b is smaller at Q4_K_M: 1.39 GiB against 2.33 GiB.

Which uses less memory at long context?

gemma-2b — its KV cache at 32K context is 8.0x smaller, because of how its attention layers cache.

Model comparison

gemma-2b vs Qwen3-4B

At Q4_K_M, gemma-2b is the smaller download — 1,495,245,728 bytes against 2,497,280,256. At long context the gap widens: gemma-2b's KV cache at 32K is 8.0× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	gemma-2b	Qwen3-4B
Parameters	2.5B	4.0B
Architecture	gemma	qwen3
Layers	18	36
Native context	8,192	40,960
Mixture of experts	no	no
Quantizations published	38	29
Smallest quantization	0.84 GiB	1.01 GiB
Q4_K_M	1.39 GiB	2.33 GiB
Licence	cc-by-nc-4.0	—

KV cache by context

the term that decides long-context viability

Context	gemma-2b	Qwen3-4B	Ratio
4,096	0.07 GiB	0.56 GiB	8.00×
8,192	0.14 GiB	1.13 GiB	8.00×
16,384	0.28 GiB	2.25 GiB	8.00×
32,768	0.56 GiB	4.50 GiB	8.00×
65,536	1.13 GiB	9.00 GiB	8.00×
131,072	2.25 GiB	18.00 GiB	8.00×