Is Qwen3.6-35B-A3B or Qwen3-VL-32B-Instruct-ultra-uncensored-heretic smaller?

Qwen3-VL-32B-Instruct-ultra-uncensored-heretic is smaller at Q4_K_M: 18.40 GiB against 19.71 GiB.

Which uses less memory at long context?

Qwen3.6-35B-A3B — its KV cache at 32K context is 12.8x smaller, because of how its attention layers cache.

Model comparison

Qwen3.6-35B-A3B vs Qwen3-VL-32B-Instruct-ultra-uncensored-heretic

At Q4_K_M, Qwen3-VL-32B-Instruct-ultra-uncensored-heretic is the smaller download — 19,762,151,392 bytes against 21,166,757,728. At long context the gap widens: Qwen3.6-35B-A3B's KV cache at 32K is 12.8× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Qwen3.6-35B-A3B	Qwen3-VL-32B-Instruct-ultra-uncensored-heretic
Parameters	36.0B	33.4B
Architecture	qwen35moe	qwen3vl
Layers	40	64
Native context	262,144	262,144
Mixture of experts	yes, 256 experts	no
Quantizations published	61	43
Smallest quantization	8.77 GiB	6.82 GiB
Q4_K_M	19.71 GiB	18.40 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	Qwen3.6-35B-A3B	Qwen3-VL-32B-Instruct-ultra-uncensored-heretic	Ratio
4,096	0.08 GiB	1.00 GiB	12.79×
8,192	0.16 GiB	2.00 GiB	12.80×
16,384	0.31 GiB	4.00 GiB	12.80×
32,768	0.63 GiB	8.00 GiB	12.80×
65,536	1.25 GiB	16.00 GiB	12.80×
131,072	2.50 GiB	32.00 GiB	12.80×