Model comparison

DarkIdol-Llama-3.1-8B-Instruct-1.3-Uncensored vs Qwen3-VL-8B-Instruct-abliterated-v1

At Q4_K_M, DarkIdol-Llama-3.1-8B-Instruct-1.3-Uncensored is the smaller download — 4,920,735,680 bytes against 5,027,785,792.

From the file· summed bytes, KV per layer

Side by side

DarkIdol-Llama-3.1-8B-Instruct-1.3-UncensoredQwen3-VL-8B-Instruct-abliterated-v1
Parameters8.0B8.8B
Architecturellamaqwen3vl
Layers32
Native context131,072
Mixture of expertsnono
Quantizations published3648
Smallest quantization1.88 GiB1.97 GiB
Q4_K_M4.58 GiB4.68 GiB
Licencellama3.1apache-2.0

KV cache by context

the term that decides long-context viability
ContextDarkIdol-Llama-3.1-8B-Instruct-1.3-UncensoredQwen3-VL-8B-Instruct-abliterated-v1Ratio
4,0960.50 GiB
8,1921.00 GiB
16,3842.00 GiB
32,7684.00 GiB
65,5368.00 GiB
131,07216.00 GiB