Is codegeex4-all-9b or Qwen3-8B smaller?

Qwen3-8B is smaller at Q4_K_M: 4.68 GiB against 5.82 GiB.

Which uses less memory at long context?

Qwen3-8B — its KV cache at 32K context is 4.4x smaller, because of how its attention layers cache.

Model comparison

codegeex4-all-9b vs Qwen3-8B

At Q4_K_M, Qwen3-8B is the smaller download — 5,027,783,488 bytes against 6,250,923,136. At long context the gap widens: Qwen3-8B's KV cache at 32K is 4.4× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	codegeex4-all-9b	Qwen3-8B
Parameters	9.4B	8.2B
Architecture	chatglm	qwen3
Layers	40	36
Native context	—	40,960
Mixture of experts	no	no
Quantizations published	53	51
Smallest quantization	2.89 GiB	2.12 GiB
Q4_K_M	5.82 GiB	4.68 GiB
Licence	other	apache-2.0

KV cache by context

the term that decides long-context viability

Context	codegeex4-all-9b	Qwen3-8B	Ratio
4,096	2.50 GiB	0.56 GiB	4.44×
8,192	5.00 GiB	1.13 GiB	4.44×
16,384	10.00 GiB	2.25 GiB	4.44×
32,768	20.00 GiB	4.50 GiB	4.44×
65,536	40.00 GiB	9.00 GiB	4.44×
131,072	80.00 GiB	18.00 GiB	4.44×