Model comparison

Z-Image-Turbo vs Chroma

At Q4_K_M, Z-Image-Turbo is the smaller download — 4,981,532,736 bytes against 5,566,533,792.

From the file· summed bytes, KV per layer

Side by side

Z-Image-TurboChroma
Parameters6.2B8.9B
Architecturelumina2flux
Layers
Native context
Mixture of expertsnono
Quantizations published30273
Smallest quantization2.41 GiB3.99 GiB
Q4_K_M4.64 GiB5.18 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextZ-Image-TurboChromaRatio
4,096
8,192
16,384
32,768
65,536
131,072