Is L3.2-Rogue-Creative-Instruct-Uncensored-Abliterated-7B or Qwen3-8B smaller?

L3.2-Rogue-Creative-Instruct-Uncensored-Abliterated-7B is smaller at Q4_K_M: 4.27 GiB against 4.68 GiB.

Which uses less memory at long context?

Qwen3-8B — its KV cache at 32K context is 1.9x smaller, because of how its attention layers cache.

Model comparison

L3.2-Rogue-Creative-Instruct-Uncensored-Abliterated-7B vs Qwen3-8B

At Q4_K_M, L3.2-Rogue-Creative-Instruct-Uncensored-Abliterated-7B is the smaller download — 4,588,966,592 bytes against 5,027,783,488. At long context the gap widens: Qwen3-8B's KV cache at 32K is 1.9× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	L3.2-Rogue-Creative-Instruct-Uncensored-Abliterated-7B	Qwen3-8B
Parameters	7.5B	8.2B
Architecture	llama	qwen3
Layers	67	36
Native context	131,072	40,960
Mixture of experts	no	no
Quantizations published	10	51
Smallest quantization	2.73 GiB	2.12 GiB
Q4_K_M	4.27 GiB	4.68 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	L3.2-Rogue-Creative-Instruct-Uncensored-Abliterated-7B	Qwen3-8B	Ratio
4,096	1.05 GiB	0.56 GiB	1.86×
8,192	2.09 GiB	1.13 GiB	1.86×
16,384	4.19 GiB	2.25 GiB	1.86×
32,768	8.38 GiB	4.50 GiB	1.86×
65,536	16.75 GiB	9.00 GiB	1.86×
131,072	33.50 GiB	18.00 GiB	1.86×