Is gemma-4-12B-it-qat-q4_0-unquantized or Kimi-Linear-48B-A3B-Instruct smaller?

Both publish quantizations; compare the table for the exact sizes.

Which uses less memory at long context?

Kimi-Linear-48B-A3B-Instruct — its KV cache at 32K context is 2.6x smaller, because of how its attention layers cache.

Model comparison

gemma-4-12B-it-qat-q4_0-unquantized vs Kimi-Linear-48B-A3B-Instruct

These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: Kimi-Linear-48B-A3B-Instruct's KV cache at 32K is 2.6× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	gemma-4-12B-it-qat-q4_0-unquantized	Kimi-Linear-48B-A3B-Instruct
Parameters	12.0B	49.1B
Architecture	gemma4	kimi-linear
Layers	48	27
Native context	262,144	—
Mixture of experts	no	yes, 256 experts
Quantizations published	2	38
Smallest quantization	6.50 GiB	9.77 GiB
Q4_K_M	—	27.66 GiB
Licence	apache-2.0	mit

KV cache by context

the term that decides long-context viability

Context	gemma-4-12B-it-qat-q4_0-unquantized	Kimi-Linear-48B-A3B-Instruct	Ratio
4,096	0.72 GiB	0.12 GiB	6.06×
8,192	0.97 GiB	0.24 GiB	4.08×
16,384	1.47 GiB	0.47 GiB	3.09×
32,768	2.47 GiB	0.95 GiB	2.60×
65,536	4.47 GiB	1.90 GiB	2.35×
131,072	8.47 GiB	3.80 GiB	2.23×