Is Qwen3.6-27B or internlm2-math-plus-20b smaller?

internlm2-math-plus-20b is smaller at Q4_K_M: 11.16 GiB against 15.41 GiB.

Which uses less memory at long context?

Qwen3.6-27B — its KV cache at 32K context is 3.0x smaller, because of how its attention layers cache.

Model comparison

Qwen3.6-27B vs internlm2-math-plus-20b

At Q4_K_M, internlm2-math-plus-20b is the smaller download — 11,984,469,440 bytes against 16,547,397,888. At long context the gap widens: Qwen3.6-27B's KV cache at 32K is 3.0× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Qwen3.6-27B	internlm2-math-plus-20b
Parameters	27.8B	19.9B
Architecture	qwen35	internlm2
Layers	64	48
Native context	262,144	8,192
Mixture of experts	no	no
Quantizations published	40	68
Smallest quantization	8.74 GiB	4.23 GiB
Q4_K_M	15.41 GiB	11.16 GiB
Licence	apache-2.0	other

KV cache by context

the term that decides long-context viability

Context	Qwen3.6-27B	internlm2-math-plus-20b	Ratio
4,096	0.25 GiB	0.75 GiB	3.00×
8,192	0.50 GiB	1.50 GiB	3.00×
16,384	1.00 GiB	3.00 GiB	3.00×
32,768	2.00 GiB	6.00 GiB	3.00×
65,536	4.00 GiB	12.00 GiB	3.00×
131,072	8.00 GiB	24.00 GiB	3.00×