Is deepseek-llm-67b-base or Qwen3.6-27B smaller?

Qwen3.6-27B is smaller at Q4_K_M: 15.41 GiB against 37.66 GiB.

Which uses less memory at long context?

Qwen3.6-27B — its KV cache at 32K context is 5.9x smaller, because of how its attention layers cache.

Model comparison

deepseek-llm-67b-base vs Qwen3.6-27B

At Q4_K_M, Qwen3.6-27B is the smaller download — 16,547,397,888 bytes against 40,439,686,592. At long context the gap widens: Qwen3.6-27B's KV cache at 32K is 5.9× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	deepseek-llm-67b-base	Qwen3.6-27B
Parameters	67.4B	27.8B
Architecture	llama	qwen35
Layers	95	64
Native context	4,096	262,144
Mixture of experts	no	no
Quantizations published	30	40
Smallest quantization	13.65 GiB	8.74 GiB
Q4_K_M	37.66 GiB	15.41 GiB
Licence	other	apache-2.0

KV cache by context

the term that decides long-context viability

Context	deepseek-llm-67b-base	Qwen3.6-27B	Ratio
4,096	1.48 GiB	0.25 GiB	5.93×
8,192	2.97 GiB	0.50 GiB	5.94×
16,384	5.94 GiB	1.00 GiB	5.94×
32,768	11.88 GiB	2.00 GiB	5.94×
65,536	23.75 GiB	4.00 GiB	5.94×
131,072	47.50 GiB	8.00 GiB	5.94×