Is Qwen3-Coder-30B-A3B-Instruct or llama2-22b-chat-wizard-uncensored 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 10.8x smaller, because of how its attention layers cache.

Model comparison

Qwen3-Coder-30B-A3B-Instruct vs llama2-22b-chat-wizard-uncensored

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 10.8× 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	llama2-22b-chat-wizard-uncensored
Parameters	30.5B	21.8B
Architecture	qwen3moe	llama
Layers	48	40
Native context	262,144	2,048
Mixture of experts	yes, 128 experts	no
Quantizations published	46	2
Smallest quantization	7.46 GiB	7.56 GiB
Q4_K_M	17.28 GiB	—
Licence	apache-2.0	other

KV cache by context

the term that decides long-context viability

Context	Qwen3-Coder-30B-A3B-Instruct	llama2-22b-chat-wizard-uncensored	Ratio
4,096	0.38 GiB	4.06 GiB	10.83×
8,192	0.75 GiB	8.13 GiB	10.83×
16,384	1.50 GiB	16.25 GiB	10.83×
32,768	3.00 GiB	32.50 GiB	10.83×
65,536	6.00 GiB	65.00 GiB	10.83×
131,072	12.00 GiB	130.00 GiB	10.83×