Is Phi-3-mini-4k-instruct or Qwen3-1.7B smaller?

Qwen3-1.7B is smaller at Q4_K_M: 1.03 GiB against 2.23 GiB.

Which uses less memory at long context?

Qwen3-1.7B — its KV cache at 32K context is 3.4x smaller, because of how its attention layers cache.

Model comparison

Phi-3-mini-4k-instruct vs Qwen3-1.7B

At Q4_K_M, Qwen3-1.7B is the smaller download — 1,107,409,472 bytes against 2,393,231,456. At long context the gap widens: Qwen3-1.7B's KV cache at 32K is 3.4× 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	Qwen3-1.7B
Parameters	3.8B	2.0B
Architecture	phi3	qwen3
Layers	32	28
Native context	4,096	40,960
Mixture of experts	no	no
Quantizations published	68	48
Smallest quantization	0.78 GiB	0.50 GiB
Q4_K_M	2.23 GiB	1.03 GiB
Licence	mit	—

KV cache by context

the term that decides long-context viability

Context	Phi-3-mini-4k-instruct	Qwen3-1.7B	Ratio
4,096	1.50 GiB	0.44 GiB	3.43×
8,192	3.00 GiB	0.88 GiB	3.43×
16,384	6.00 GiB	1.75 GiB	3.43×
32,768	12.00 GiB	3.50 GiB	3.43×
65,536	24.00 GiB	7.00 GiB	3.43×
131,072	48.00 GiB	14.00 GiB	3.43×