Is AMALIA-9B-0626-DPO or Qwen3-8B smaller?

Qwen3-8B is smaller at Q4_K_M: 4.68 GiB against 5.20 GiB.

Which uses less memory at long context?

Qwen3-8B — its KV cache at 32K context is 1.2x smaller, because of how its attention layers cache.

Model comparison

AMALIA-9B-0626-DPO vs Qwen3-8B

At Q4_K_M, Qwen3-8B is the smaller download — 5,027,783,488 bytes against 5,582,838,528. At long context the gap widens: Qwen3-8B's KV cache at 32K is 1.2× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	AMALIA-9B-0626-DPO	Qwen3-8B
Parameters	9.2B	8.2B
Architecture	llama	qwen3
Layers	42	36
Native context	32,768	40,960
Mixture of experts	no	no
Quantizations published	19	51
Smallest quantization	3.35 GiB	2.12 GiB
Q4_K_M	5.20 GiB	4.68 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	AMALIA-9B-0626-DPO	Qwen3-8B	Ratio
4,096	0.66 GiB	0.56 GiB	1.17×
8,192	1.31 GiB	1.13 GiB	1.17×
16,384	2.63 GiB	2.25 GiB	1.17×
32,768	5.25 GiB	4.50 GiB	1.17×
65,536	10.50 GiB	9.00 GiB	1.17×
131,072	21.00 GiB	18.00 GiB	1.17×