Is Yi-1.5-34B 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 3.0x smaller, because of how its attention layers cache.

Model comparison

Yi-1.5-34B 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 3.0× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Yi-1.5-34B	gemma-4-12B-it-qat-q4_0-unquantized
Parameters	34.4B	12.0B
Architecture	llama	gemma4
Layers	60	48
Native context	4,096	262,144
Mixture of experts	no	no
Quantizations published	37	2
Smallest quantization	6.98 GiB	6.50 GiB
Q4_K_M	19.24 GiB	—
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	Yi-1.5-34B	gemma-4-12B-it-qat-q4_0-unquantized	Ratio
4,096	0.94 GiB	0.72 GiB	1.30×
8,192	1.88 GiB	0.97 GiB	1.94×
16,384	3.75 GiB	1.47 GiB	2.55×
32,768	7.50 GiB	2.47 GiB	3.04×
65,536	15.00 GiB	4.47 GiB	3.36×
131,072	30.00 GiB	8.47 GiB	3.54×