Is Qwen3-Coder-30B-A3B-Instruct or Laguna-XS-2.1 smaller?

Qwen3-Coder-30B-A3B-Instruct is smaller at Q4_K_M: 17.28 GiB against 18.88 GiB.

Which uses less memory at long context?

Laguna-XS-2.1 — 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 Laguna-XS-2.1

At Q4_K_M, Qwen3-Coder-30B-A3B-Instruct is the smaller download — 18,556,689,568 bytes against 20,274,299,936. At long context the gap widens: Laguna-XS-2.1'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	Laguna-XS-2.1
Parameters	30.5B	33.4B
Architecture	qwen3moe	laguna
Layers	48	40
Native context	262,144	262,144
Mixture of experts	yes, 128 experts	yes, 256 experts
Quantizations published	46	31
Smallest quantization	7.46 GiB	8.76 GiB
Q4_K_M	17.28 GiB	18.88 GiB
Licence	apache-2.0	openmdw-1.1

KV cache by context

the term that decides long-context viability

Context	Qwen3-Coder-30B-A3B-Instruct	Laguna-XS-2.1	Ratio
4,096	0.38 GiB	0.27 GiB	1.37×
8,192	0.75 GiB	0.43 GiB	1.75×
16,384	1.50 GiB	0.74 GiB	2.02×
32,768	3.00 GiB	1.37 GiB	2.19×
65,536	6.00 GiB	2.62 GiB	2.29×
131,072	12.00 GiB	5.12 GiB	2.35×