Model comparison

Fimbulvetr-11B-v2 vs Qwen3-VL-8B-Instruct-abliterated-v1

At Q4_K_M, Qwen3-VL-8B-Instruct-abliterated-v1 is the smaller download — 5,027,785,792 bytes against 6,616,291,424.

From the file· summed bytes, KV per layer

Side by side

Fimbulvetr-11B-v2Qwen3-VL-8B-Instruct-abliterated-v1
Parameters10.7B8.8B
Architecturellamaqwen3vl
Layers48
Native context4,096
Mixture of expertsnono
Quantizations published3648
Smallest quantization2.35 GiB1.97 GiB
Q4_K_M6.16 GiB4.68 GiB
Licencecc-by-nc-4.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextFimbulvetr-11B-v2Qwen3-VL-8B-Instruct-abliterated-v1Ratio
4,0960.75 GiB
8,1921.50 GiB
16,3843.00 GiB
32,7686.00 GiB
65,53612.00 GiB
131,07224.00 GiB