Is Qwen3-Coder-30B-A3B-Instruct or GLM-Z1-32B-0414-uncensored-heretic-v2 smaller?

Qwen3-Coder-30B-A3B-Instruct is smaller at Q4_K_M: 17.28 GiB against 18.33 GiB.

Which uses less memory at long context?

GLM-Z1-32B-0414-uncensored-heretic-v2 — its KV cache at 32K context is 1.6x smaller, because of how its attention layers cache.

Model comparison

Qwen3-Coder-30B-A3B-Instruct vs GLM-Z1-32B-0414-uncensored-heretic-v2

At Q4_K_M, Qwen3-Coder-30B-A3B-Instruct is the smaller download — 18,556,689,568 bytes against 19,686,117,248. At long context the gap widens: GLM-Z1-32B-0414-uncensored-heretic-v2's KV cache at 32K is 1.6× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Qwen3-Coder-30B-A3B-Instruct	GLM-Z1-32B-0414-uncensored-heretic-v2
Parameters	30.5B	32.6B
Architecture	qwen3moe	glm4
Layers	48	61
Native context	262,144	32,768
Mixture of experts	yes, 128 experts	no
Quantizations published	46	9
Smallest quantization	7.46 GiB	14.80 GiB
Q4_K_M	17.28 GiB	18.33 GiB
Licence	apache-2.0	mit

KV cache by context

the term that decides long-context viability

Context	Qwen3-Coder-30B-A3B-Instruct	GLM-Z1-32B-0414-uncensored-heretic-v2	Ratio
4,096	0.38 GiB	0.24 GiB	1.57×
8,192	0.75 GiB	0.48 GiB	1.57×
16,384	1.50 GiB	0.95 GiB	1.57×
32,768	3.00 GiB	1.91 GiB	1.57×
65,536	6.00 GiB	3.81 GiB	1.57×
131,072	12.00 GiB	7.63 GiB	1.57×