Model comparison

granite-4.1-8b 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 5,347,914,400.

From the file· summed bytes, KV per layer

Side by side

granite-4.1-8bQwen3-VL-8B-Instruct-abliterated-v1
Parameters8.8B8.8B
Architectureqwen3vl
Layers40
Native context131,072
Mixture of expertsnono
Quantizations published3048
Smallest quantization3.05 GiB1.97 GiB
Q4_K_M4.98 GiB4.68 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextgranite-4.1-8bQwen3-VL-8B-Instruct-abliterated-v1Ratio
4,0960.63 GiB
8,1921.25 GiB
16,3842.50 GiB
32,7685.00 GiB
65,53610.00 GiB
131,07220.00 GiB