Is Qwen3-Coder-30B-A3B-Instruct or magnum-v4-27b smaller?

magnum-v4-27b is smaller at Q4_K_M: 15.50 GiB against 17.28 GiB.

Which uses less memory at long context?

Qwen3-Coder-30B-A3B-Instruct — its KV cache at 32K context is 2.2x smaller, because of how its attention layers cache.

Model comparison

Qwen3-Coder-30B-A3B-Instruct vs magnum-v4-27b

At Q4_K_M, magnum-v4-27b is the smaller download — 16,645,381,728 bytes against 18,556,689,568. At long context the gap widens: Qwen3-Coder-30B-A3B-Instruct's KV cache at 32K is 2.2× 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	magnum-v4-27b
Parameters	30.5B	27.2B
Architecture	qwen3moe	gemma2
Layers	48	46
Native context	262,144	8,192
Mixture of experts	yes, 128 experts	no
Quantizations published	46	22
Smallest quantization	7.46 GiB	7.82 GiB
Q4_K_M	17.28 GiB	15.50 GiB
Licence	apache-2.0	gemma

KV cache by context

the term that decides long-context viability

Context	Qwen3-Coder-30B-A3B-Instruct	magnum-v4-27b	Ratio
4,096	0.38 GiB	1.44 GiB	3.83×
8,192	0.75 GiB	2.25 GiB	2.99×
16,384	1.50 GiB	3.68 GiB	2.46×
32,768	3.00 GiB	6.56 GiB	2.19×
65,536	6.00 GiB	12.31 GiB	2.05×
131,072	12.00 GiB	23.81 GiB	1.98×