Is GLM-Z1-Rumination-32B-0414 or gemma-4-12B-it-qat-q4_0-unquantized smaller?

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

Which uses less memory at long context?

gemma-4-12B-it-qat-q4_0-unquantized — its KV cache at 32K context is 3.1x smaller, because of how its attention layers cache.

Model comparison

GLM-Z1-Rumination-32B-0414 vs gemma-4-12B-it-qat-q4_0-unquantized

These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: gemma-4-12B-it-qat-q4_0-unquantized's KV cache at 32K is 3.1× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	GLM-Z1-Rumination-32B-0414	gemma-4-12B-it-qat-q4_0-unquantized
Parameters	33.1B	12.0B
Architecture	glm4	gemma4
Layers	61	48
Native context	131,072	262,144
Mixture of experts	no	no
Quantizations published	29	2
Smallest quantization	9.45 GiB	6.50 GiB
Q4_K_M	18.66 GiB	—
Licence	mit	apache-2.0

KV cache by context

the term that decides long-context viability

Context	GLM-Z1-Rumination-32B-0414	gemma-4-12B-it-qat-q4_0-unquantized	Ratio
4,096	0.95 GiB	0.72 GiB	1.33×
8,192	1.91 GiB	0.97 GiB	1.97×
16,384	3.81 GiB	1.47 GiB	2.60×
32,768	7.63 GiB	2.47 GiB	3.09×
65,536	15.25 GiB	4.47 GiB	3.41×
131,072	30.50 GiB	8.47 GiB	3.60×