Model comparison

FLUX.1-Fill-dev vs Z-Image-Turbo

These two publish different quantization sets; the table below has the exact sizes.

From the file· summed bytes, KV per layer

Side by side

FLUX.1-Fill-devZ-Image-Turbo
Parameters11.9B6.2B
Architecturefluxlumina2
Layers
Native context
Mixture of expertsnono
Quantizations published2530
Smallest quantization3.87 GiB2.41 GiB
Q4_K_M4.64 GiB
Licenceotherapache-2.0

KV cache by context

the term that decides long-context viability
ContextFLUX.1-Fill-devZ-Image-TurboRatio
4,096
8,192
16,384
32,768
65,536
131,072