Is Phi-3-mini-4k-instruct or gemma-3-1b-it smaller?

gemma-3-1b-it is smaller at Q4_K_M: 0.75 GiB against 2.23 GiB.

Which uses less memory at long context?

gemma-3-1b-it — its KV cache at 32K context is 81.9x smaller, because of how its attention layers cache.

Model comparison

Phi-3-mini-4k-instruct vs gemma-3-1b-it

At Q4_K_M, gemma-3-1b-it is the smaller download — 806,058,240 bytes against 2,393,231,456. At long context the gap widens: gemma-3-1b-it's KV cache at 32K is 81.9× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Phi-3-mini-4k-instruct	gemma-3-1b-it
Parameters	3.8B	1000M
Architecture	phi3	gemma3
Layers	32	26
Native context	4,096	32,768
Mixture of experts	no	no
Quantizations published	68	28
Smallest quantization	0.78 GiB	0.52 GiB
Q4_K_M	2.23 GiB	0.75 GiB
Licence	mit	gemma

KV cache by context

the term that decides long-context viability

Context	Phi-3-mini-4k-instruct	gemma-3-1b-it	Ratio
4,096	1.50 GiB	0.04 GiB	40.42×
8,192	3.00 GiB	0.05 GiB	56.89×
16,384	6.00 GiB	0.08 GiB	71.44×
32,768	12.00 GiB	0.15 GiB	81.92×
65,536	24.00 GiB	0.27 GiB	88.40×
131,072	48.00 GiB	0.52 GiB	92.04×