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

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

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

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

KV cache by context

the term that decides long-context viability

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