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

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

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

	llama2-22b-chat-wizard-uncensored	Qwen3-Coder-30B-A3B-Instruct
Parameters	21.8B	30.5B
Architecture	llama	qwen3moe
Layers	40	48
Native context	2,048	262,144
Mixture of experts	no	yes, 128 experts
Quantizations published	2	46
Smallest quantization	7.56 GiB	7.46 GiB
Q4_K_M	—	17.28 GiB
Licence	other	apache-2.0

KV cache by context

the term that decides long-context viability

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