Is AfriqueQwen3.5-4B or Qwen3-VL-4B-Instruct smaller?

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

Which uses less memory at long context?

AfriqueQwen3.5-4B — its KV cache at 32K context is 4.5x smaller, because of how its attention layers cache.

Model comparison

AfriqueQwen3.5-4B vs Qwen3-VL-4B-Instruct

These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: AfriqueQwen3.5-4B's KV cache at 32K is 4.5× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	AfriqueQwen3.5-4B	Qwen3-VL-4B-Instruct
Parameters	5.2B	4.4B
Architecture	qwen35	qwen3vl
Layers	32	36
Native context	262,144	262,144
Mixture of experts	no	no
Quantizations published	24	26
Smallest quantization	1.46 GiB	1.01 GiB
Q4_K_M	—	2.33 GiB
Licence	cc-by-4.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	AfriqueQwen3.5-4B	Qwen3-VL-4B-Instruct	Ratio
4,096	0.13 GiB	0.56 GiB	4.50×
8,192	0.25 GiB	1.13 GiB	4.50×
16,384	0.50 GiB	2.25 GiB	4.50×
32,768	1.00 GiB	4.50 GiB	4.50×
65,536	2.00 GiB	9.00 GiB	4.50×
131,072	4.00 GiB	18.00 GiB	4.50×