Is phi-2 or Qwen3-8B smaller?

phi-2 is smaller at Q4_K_M: 1.62 GiB against 4.68 GiB.

Which uses less memory at long context?

Qwen3-8B — its KV cache at 32K context is 2.2x smaller, because of how its attention layers cache.

Model comparison

phi-2 vs Qwen3-8B

At Q4_K_M, phi-2 is the smaller download — 1,737,636,384 bytes against 5,027,783,488. At long context the gap widens: Qwen3-8B's KV cache at 32K is 2.2× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	phi-2	Qwen3-8B
Parameters	2.8B	8.2B
Architecture	phi2	qwen3
Layers	32	36
Native context	2,048	40,960
Mixture of experts	no	no
Quantizations published	34	51
Smallest quantization	1.03 GiB	2.12 GiB
Q4_K_M	1.62 GiB	4.68 GiB
Licence	other	apache-2.0

KV cache by context

the term that decides long-context viability

Context	phi-2	Qwen3-8B	Ratio
4,096	1.25 GiB	0.56 GiB	2.22×
8,192	2.50 GiB	1.13 GiB	2.22×
16,384	5.00 GiB	2.25 GiB	2.22×
32,768	10.00 GiB	4.50 GiB	2.22×
65,536	20.00 GiB	9.00 GiB	2.22×
131,072	40.00 GiB	18.00 GiB	2.22×