Is Qwen3.5-9B or Ministral-3-8B-Reasoning-2512-heretic smaller?

Ministral-3-8B-Reasoning-2512-heretic is smaller at Q4_K_M: 4.84 GiB against 5.24 GiB.

Which uses less memory at long context?

Qwen3.5-9B — its KV cache at 32K context is 4.2x smaller, because of how its attention layers cache.

Model comparison

Qwen3.5-9B vs Ministral-3-8B-Reasoning-2512-heretic

At Q4_K_M, Ministral-3-8B-Reasoning-2512-heretic is the smaller download — 5,198,385,440 bytes against 5,627,044,256. At long context the gap widens: Qwen3.5-9B's KV cache at 32K is 4.2× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Qwen3.5-9B	Ministral-3-8B-Reasoning-2512-heretic
Parameters	9.7B	8.9B
Architecture	qwen35	mistral3
Layers	32	34
Native context	262,144	262,144
Mixture of experts	no	no
Quantizations published	34	3
Smallest quantization	2.97 GiB	4.84 GiB
Q4_K_M	5.24 GiB	4.84 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	Qwen3.5-9B	Ministral-3-8B-Reasoning-2512-heretic	Ratio
4,096	0.13 GiB	0.53 GiB	4.25×
8,192	0.25 GiB	1.06 GiB	4.25×
16,384	0.50 GiB	2.13 GiB	4.25×
32,768	1.00 GiB	4.25 GiB	4.25×
65,536	2.00 GiB	8.50 GiB	4.25×
131,072	4.00 GiB	17.00 GiB	4.25×