Is Qwen2-VL-2B-Instruct or LFM2.5-VL-1.6B smaller?

LFM2.5-VL-1.6B is smaller at Q4_K_M: 0.68 GiB against 0.92 GiB.

Which uses less memory at long context?

LFM2.5-VL-1.6B — its KV cache at 32K context is 2.3x smaller, because of how its attention layers cache.

Model comparison

Qwen2-VL-2B-Instruct vs LFM2.5-VL-1.6B

At Q4_K_M, LFM2.5-VL-1.6B is the smaller download — 730,895,968 bytes against 986,046,944. At long context the gap widens: LFM2.5-VL-1.6B's KV cache at 32K is 2.3× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Qwen2-VL-2B-Instruct	LFM2.5-VL-1.6B
Parameters	2.2B	1.6B
Architecture	qwen2vl	lfm2
Layers	28	16
Native context	32,768	128,000
Mixture of experts	no	no
Quantizations published	25	23
Smallest quantization	0.56 GiB	0.45 GiB
Q4_K_M	0.92 GiB	0.68 GiB
Licence	apache-2.0	other

KV cache by context

the term that decides long-context viability

Context	Qwen2-VL-2B-Instruct	LFM2.5-VL-1.6B	Ratio
4,096	0.11 GiB	0.05 GiB	2.33×
8,192	0.22 GiB	0.09 GiB	2.33×
16,384	0.44 GiB	0.19 GiB	2.33×
32,768	0.88 GiB	0.38 GiB	2.33×
65,536	1.75 GiB	0.75 GiB	2.33×
131,072	3.50 GiB	1.50 GiB	2.33×