Is Qwen3-Coder-30B-A3B-Instruct or dolphin-2.9.1-mixtral-1x22b smaller?

Both publish quantizations; compare the table for the exact sizes.

Which uses less memory at long context?

Qwen3-Coder-30B-A3B-Instruct — its KV cache at 32K context is 2.3x smaller, because of how its attention layers cache.

Model comparison

Qwen3-Coder-30B-A3B-Instruct vs dolphin-2.9.1-mixtral-1x22b

These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: Qwen3-Coder-30B-A3B-Instruct's KV cache at 32K is 2.3× 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	dolphin-2.9.1-mixtral-1x22b
Parameters	30.5B	22.2B
Architecture	qwen3moe	llama
Layers	48	56
Native context	262,144	65,536
Mixture of experts	yes, 128 experts	yes, 1 experts
Quantizations published	46	45
Smallest quantization	7.46 GiB	4.49 GiB
Q4_K_M	17.28 GiB	—
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	Qwen3-Coder-30B-A3B-Instruct	dolphin-2.9.1-mixtral-1x22b	Ratio
4,096	0.38 GiB	0.88 GiB	2.33×
8,192	0.75 GiB	1.75 GiB	2.33×
16,384	1.50 GiB	3.50 GiB	2.33×
32,768	3.00 GiB	7.00 GiB	2.33×
65,536	6.00 GiB	14.00 GiB	2.33×
131,072	12.00 GiB	28.00 GiB	2.33×