Is gemma-4-26B-A4B-it-assistant or Qwen3-1.7B smaller?

gemma-4-26B-A4B-it-assistant is smaller at Q4_K_M: 0.30 GiB against 1.03 GiB.

Which uses less memory at long context?

gemma-4-26B-A4B-it-assistant — its KV cache at 32K context is 12.3x smaller, because of how its attention layers cache.

Model comparison

gemma-4-26B-A4B-it-assistant vs Qwen3-1.7B

At Q4_K_M, gemma-4-26B-A4B-it-assistant is the smaller download — 325,452,000 bytes against 1,107,409,472. At long context the gap widens: gemma-4-26B-A4B-it-assistant's KV cache at 32K is 12.3× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	gemma-4-26B-A4B-it-assistant	Qwen3-1.7B
Parameters	420M	2.0B
Architecture	gemma4-assistant	qwen3
Layers	4	28
Native context	262,144	40,960
Mixture of experts	no	no
Quantizations published	14	48
Smallest quantization	0.30 GiB	0.50 GiB
Q4_K_M	0.30 GiB	1.03 GiB
Licence	apache-2.0	—

KV cache by context

the term that decides long-context viability

Context	gemma-4-26B-A4B-it-assistant	Qwen3-1.7B	Ratio
4,096	0.07 GiB	0.44 GiB	6.59×
8,192	0.10 GiB	0.88 GiB	8.96×
16,384	0.16 GiB	1.75 GiB	10.93×
32,768	0.29 GiB	3.50 GiB	12.27×
65,536	0.54 GiB	7.00 GiB	13.08×
131,072	1.04 GiB	14.00 GiB	13.52×