Is Qwen3.5-0.8B-heretic or Qwen2-VL-2B-Instruct smaller?

Qwen3.5-0.8B-heretic is smaller at Q4_K_M: 0.49 GiB against 0.92 GiB.

Which uses less memory at long context?

Qwen3.5-0.8B-heretic — its KV cache at 32K context is 2.3x smaller, because of how its attention layers cache.

Model comparison

Qwen3.5-0.8B-heretic vs Qwen2-VL-2B-Instruct

At Q4_K_M, Qwen3.5-0.8B-heretic is the smaller download — 529,297,312 bytes against 986,046,944. At long context the gap widens: Qwen3.5-0.8B-heretic'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

	Qwen3.5-0.8B-heretic	Qwen2-VL-2B-Instruct
Parameters	853M	2.2B
Architecture	clip	qwen2vl
Layers	24	28
Native context	262,144	32,768
Mixture of experts	no	no
Quantizations published	6	25
Smallest quantization	0.26 GiB	0.56 GiB
Q4_K_M	0.49 GiB	0.92 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	Qwen3.5-0.8B-heretic	Qwen2-VL-2B-Instruct	Ratio
4,096	0.05 GiB	0.11 GiB	2.33×
8,192	0.09 GiB	0.22 GiB	2.33×
16,384	0.19 GiB	0.44 GiB	2.33×
32,768	0.38 GiB	0.88 GiB	2.33×
65,536	0.75 GiB	1.75 GiB	2.33×
131,072	1.50 GiB	3.50 GiB	2.33×