Model comparison

AuraFlow-v0.3 vs FLUX.1-dev

At Q4_K_M, AuraFlow-v0.3 is the smaller download — 4,047,070,496 bytes against 6,931,817,760.

From the file· summed bytes, KV per layer

Side by side

AuraFlow-v0.3FLUX.1-dev
Parameters6.9B11.9B
Architectureauraflux
Layers
Native context
Mixture of expertsnono
Quantizations published1437
Smallest quantization2.23 GiB3.75 GiB
Q4_K_M3.77 GiB6.46 GiB
Licenceapache-2.0other

KV cache by context

the term that decides long-context viability
ContextAuraFlow-v0.3FLUX.1-devRatio
4,096
8,192
16,384
32,768
65,536
131,072