Is Huihui-Qwen3-VL-32B-Instruct-abliterated or gemma-4-E4B-it smaller?

gemma-4-E4B-it is smaller at Q4_K_M: 4.64 GiB against 18.40 GiB.

Which uses less memory at long context?

gemma-4-E4B-it — its KV cache at 32K context is 15.8x smaller, because of how its attention layers cache.

Model comparison

Huihui-Qwen3-VL-32B-Instruct-abliterated vs gemma-4-E4B-it

At Q4_K_M, gemma-4-E4B-it is the smaller download — 4,977,171,584 bytes against 19,762,151,264. At long context the gap widens: gemma-4-E4B-it's KV cache at 32K is 15.8× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Huihui-Qwen3-VL-32B-Instruct-abliterated	gemma-4-E4B-it
Parameters	33.4B	8.0B
Architecture	qwen3vl	gemma4
Layers	64	42
Native context	262,144	131,072
Mixture of experts	no	no
Quantizations published	34	36
Smallest quantization	6.82 GiB	3.30 GiB
Q4_K_M	18.40 GiB	4.64 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	Huihui-Qwen3-VL-32B-Instruct-abliterated	gemma-4-E4B-it	Ratio
4,096	1.00 GiB	0.12 GiB	8.13×
8,192	2.00 GiB	0.18 GiB	11.25×
16,384	4.00 GiB	0.29 GiB	13.93×
32,768	8.00 GiB	0.51 GiB	15.81×
65,536	16.00 GiB	0.94 GiB	16.96×
131,072	32.00 GiB	1.82 GiB	17.60×