Is Magistral-Small-2509-Vision or Qwen3-VL-30B-A3B-Instruct smaller?

Magistral-Small-2509-Vision is smaller at Q4_K_M: 11.76 GiB against 17.28 GiB.

Which uses less memory at long context?

Qwen3-VL-30B-A3B-Instruct — its KV cache at 32K context is 1.7x smaller, because of how its attention layers cache.

Model comparison

Magistral-Small-2509-Vision vs Qwen3-VL-30B-A3B-Instruct

At Q4_K_M, Magistral-Small-2509-Vision is the smaller download — 12,621,946,560 bytes against 18,556,687,168. At long context the gap widens: Qwen3-VL-30B-A3B-Instruct's KV cache at 32K is 1.7× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Magistral-Small-2509-Vision	Qwen3-VL-30B-A3B-Instruct
Parameters	24.0B	31.1B
Architecture	llama	qwen3vlmoe
Layers	40	48
Native context	131,072	262,144
Mixture of experts	no	yes, 128 experts
Quantizations published	16	26
Smallest quantization	8.09 GiB	7.60 GiB
Q4_K_M	11.76 GiB	17.28 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	Magistral-Small-2509-Vision	Qwen3-VL-30B-A3B-Instruct	Ratio
4,096	0.63 GiB	0.38 GiB	1.67×
8,192	1.25 GiB	0.75 GiB	1.67×
16,384	2.50 GiB	1.50 GiB	1.67×
32,768	5.00 GiB	3.00 GiB	1.67×
65,536	10.00 GiB	6.00 GiB	1.67×
131,072	20.00 GiB	12.00 GiB	1.67×