Model comparison

Illustrious-xl-early-release-v0 vs Z-Image-Turbo

At Q4_K_M, Z-Image-Turbo is the smaller download — 4,981,532,736 bytes against 6,109,905,792.

From the file· summed bytes, KV per layer

Side by side

Illustrious-xl-early-release-v0Z-Image-Turbo
Parameters2.6B6.2B
Architecturesdxllumina2
Layers
Native context
Mixture of expertsnono
Quantizations published1530
Smallest quantization3.30 GiB2.41 GiB
Q4_K_M5.69 GiB4.64 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextIllustrious-xl-early-release-v0Z-Image-TurboRatio
4,096
8,192
16,384
32,768
65,536
131,072