Is archangel_sft-kto_llama30b or Qwen3.8-27B smaller?

Qwen3.8-27B is smaller at Q4_K_M: 15.66 GiB against 18.27 GiB.

Which uses less memory at long context?

Qwen3.8-27B — its KV cache at 32K context is 24.4x smaller, because of how its attention layers cache.

Model comparison

archangel_sft-kto_llama30b vs Qwen3.8-27B

At Q4_K_M, Qwen3.8-27B is the smaller download — 16,810,714,336 bytes against 19,621,140,512. At long context the gap widens: Qwen3.8-27B's KV cache at 32K is 24.4× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	archangel_sft-kto_llama30b	Qwen3.8-27B
Parameters	32.5B	27.8B
Architecture	llama	qwen35
Layers	60	64
Native context	2,048	262,144
Mixture of experts	no	no
Quantizations published	35	22
Smallest quantization	6.63 GiB	8.39 GiB
Q4_K_M	18.27 GiB	15.66 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	archangel_sft-kto_llama30b	Qwen3.8-27B	Ratio
4,096	6.09 GiB	0.25 GiB	24.36×
8,192	12.19 GiB	0.50 GiB	24.37×
16,384	24.38 GiB	1.00 GiB	24.37×
32,768	48.75 GiB	2.00 GiB	24.37×
65,536	97.50 GiB	4.00 GiB	24.37×
131,072	195.00 GiB	8.00 GiB	24.37×