Model comparison

granite-4.0-micro-base vs Qwen3-VL-8B-Instruct-abliterated-v1

At Q4_K_M, granite-4.0-micro-base is the smaller download — 2,099,496,064 bytes against 5,027,785,792.

From the file· summed bytes, KV per layer

Side by side

granite-4.0-micro-baseQwen3-VL-8B-Instruct-abliterated-v1
Parameters3.4B8.8B
Architecturegraniteqwen3vl
Layers40
Native context131,072
Mixture of expertsnono
Quantizations published1548
Smallest quantization1.28 GiB1.97 GiB
Q4_K_M1.96 GiB4.68 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextgranite-4.0-micro-baseQwen3-VL-8B-Instruct-abliterated-v1Ratio
4,0960.31 GiB
8,1920.63 GiB
16,3841.25 GiB
32,7682.50 GiB
65,5365.00 GiB
131,07210.00 GiB