Is UncensoredLM-DeepSeek-R1-Distill-Qwen-14B or Qwen3.6-27B smaller?

UncensoredLM-DeepSeek-R1-Distill-Qwen-14B is smaller at Q4_K_M: 8.05 GiB against 15.41 GiB.

Which uses less memory at long context?

Qwen3.6-27B — its KV cache at 32K context is 2.9x smaller, because of how its attention layers cache.

Model comparison

UncensoredLM-DeepSeek-R1-Distill-Qwen-14B vs Qwen3.6-27B

At Q4_K_M, UncensoredLM-DeepSeek-R1-Distill-Qwen-14B is the smaller download — 8,639,113,408 bytes against 16,547,397,888. At long context the gap widens: Qwen3.6-27B's KV cache at 32K is 2.9× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	UncensoredLM-DeepSeek-R1-Distill-Qwen-14B	Qwen3.6-27B
Parameters	14.2B	27.8B
Architecture	qwen2	qwen35
Layers	46	64
Native context	131,072	262,144
Mixture of experts	no	no
Quantizations published	25	40
Smallest quantization	4.23 GiB	8.74 GiB
Q4_K_M	8.05 GiB	15.41 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	UncensoredLM-DeepSeek-R1-Distill-Qwen-14B	Qwen3.6-27B	Ratio
4,096	0.72 GiB	0.25 GiB	2.87×
8,192	1.44 GiB	0.50 GiB	2.87×
16,384	2.88 GiB	1.00 GiB	2.87×
32,768	5.75 GiB	2.00 GiB	2.87×
65,536	11.50 GiB	4.00 GiB	2.87×
131,072	23.00 GiB	8.00 GiB	2.87×