Is gemma-4-26B-A4B-it-DFlash or Qwen2.5-0.5B-Instruct smaller?

gemma-4-26B-A4B-it-DFlash is smaller at Q4_K_M: 0.25 GiB against 0.37 GiB.

Which uses less memory at long context?

gemma-4-26B-A4B-it-DFlash — its KV cache at 32K context is 2.3x smaller, because of how its attention layers cache.

Model comparison

gemma-4-26B-A4B-it-DFlash vs Qwen2.5-0.5B-Instruct

At Q4_K_M, gemma-4-26B-A4B-it-DFlash is the smaller download — 266,221,536 bytes against 397,807,936. At long context the gap widens: gemma-4-26B-A4B-it-DFlash's KV cache at 32K is 2.3× 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-DFlash	Qwen2.5-0.5B-Instruct
Parameters	430M	494M
Architecture	dflash	qwen2
Layers	5	24
Native context	262,144	32,768
Mixture of experts	no	no
Quantizations published	18	32
Smallest quantization	0.16 GiB	0.31 GiB
Q4_K_M	0.25 GiB	0.37 GiB
Licence	—	apache-2.0

KV cache by context

the term that decides long-context viability

Context	gemma-4-26B-A4B-it-DFlash	Qwen2.5-0.5B-Instruct	Ratio
4,096	0.05 GiB	0.05 GiB	1.17×
8,192	0.07 GiB	0.09 GiB	1.33×
16,384	0.10 GiB	0.19 GiB	1.85×
32,768	0.16 GiB	0.38 GiB	2.29×
65,536	0.29 GiB	0.75 GiB	2.59×
131,072	0.54 GiB	1.50 GiB	2.78×