Is Qwen3.6-27B or Mellum2-12B-A2.5B-Thinking smaller?

Mellum2-12B-A2.5B-Thinking is smaller at Q4_K_M: 7.52 GiB against 15.41 GiB.

Which uses less memory at long context?

Mellum2-12B-A2.5B-Thinking — its KV cache at 32K context is 4.0x smaller, because of how its attention layers cache.

Model comparison

Qwen3.6-27B vs Mellum2-12B-A2.5B-Thinking

At Q4_K_M, Mellum2-12B-A2.5B-Thinking is the smaller download — 8,071,295,040 bytes against 16,547,397,888. At long context the gap widens: Mellum2-12B-A2.5B-Thinking's KV cache at 32K is 4.0× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Qwen3.6-27B	Mellum2-12B-A2.5B-Thinking
Parameters	27.8B	12.1B
Architecture	qwen35	mellum
Layers	64	28
Native context	262,144	131,072
Mixture of experts	no	yes, 64 experts
Quantizations published	40	3
Smallest quantization	8.74 GiB	7.52 GiB
Q4_K_M	15.41 GiB	7.52 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	Qwen3.6-27B	Mellum2-12B-A2.5B-Thinking	Ratio
4,096	0.25 GiB	0.12 GiB	2.15×
8,192	0.50 GiB	0.17 GiB	2.93×
16,384	1.00 GiB	0.28 GiB	3.57×
32,768	2.00 GiB	0.50 GiB	4.01×
65,536	4.00 GiB	0.94 GiB	4.27×
131,072	8.00 GiB	1.81 GiB	4.42×