Is SmolVLM-Instruct or Qwen3-8B smaller?

SmolVLM-Instruct is smaller at Q4_K_M: 1.04 GiB against 4.68 GiB.

Which uses less memory at long context?

Qwen3-8B — its KV cache at 32K context is 1.3x smaller, because of how its attention layers cache.

Model comparison

SmolVLM-Instruct vs Qwen3-8B

At Q4_K_M, SmolVLM-Instruct is the smaller download — 1,112,242,368 bytes against 5,027,783,488. At long context the gap widens: Qwen3-8B's KV cache at 32K is 1.3× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	SmolVLM-Instruct	Qwen3-8B
Parameters	2.2B	8.2B
Architecture	llama	qwen3
Layers	24	36
Native context	16,384	40,960
Mixture of experts	no	no
Quantizations published	3	51
Smallest quantization	1.04 GiB	2.12 GiB
Q4_K_M	1.04 GiB	4.68 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	SmolVLM-Instruct	Qwen3-8B	Ratio
4,096	0.75 GiB	0.56 GiB	1.33×
8,192	1.50 GiB	1.13 GiB	1.33×
16,384	3.00 GiB	2.25 GiB	1.33×
32,768	6.00 GiB	4.50 GiB	1.33×
65,536	12.00 GiB	9.00 GiB	1.33×
131,072	24.00 GiB	18.00 GiB	1.33×