Is gemma4-12B-it-DFlash or Qwen3-1.7B smaller?

gemma4-12B-it-DFlash is smaller at Q4_K_M: 0.41 GiB against 1.03 GiB.

Which uses less memory at long context?

gemma4-12B-it-DFlash — its KV cache at 32K context is 21.3x smaller, because of how its attention layers cache.

Model comparison

gemma4-12B-it-DFlash vs Qwen3-1.7B

At Q4_K_M, gemma4-12B-it-DFlash is the smaller download — 442,398,496 bytes against 1,107,409,472. At long context the gap widens: gemma4-12B-it-DFlash's KV cache at 32K is 21.3× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	gemma4-12B-it-DFlash	Qwen3-1.7B
Parameters	727M	2.0B
Architecture	dflash	qwen3
Layers	5	28
Native context	262,144	40,960
Mixture of experts	no	no
Quantizations published	6	48
Smallest quantization	0.38 GiB	0.50 GiB
Q4_K_M	0.41 GiB	1.03 GiB
Licence	apache-2.0	—

KV cache by context

the term that decides long-context viability

Context	gemma4-12B-it-DFlash	Qwen3-1.7B	Ratio
4,096	0.05 GiB	0.44 GiB	8.00×
8,192	0.07 GiB	0.88 GiB	12.44×
16,384	0.10 GiB	1.75 GiB	17.23×
32,768	0.16 GiB	3.50 GiB	21.33×
65,536	0.29 GiB	7.00 GiB	24.22×
131,072	0.54 GiB	14.00 GiB	25.97×