Model comparison

ideogram-4-fp8 vs FLUX.1-dev

At Q4_K_M, ideogram-4-fp8 is the smaller download — 5,027,785,568 bytes against 6,931,817,760.

From the file· summed bytes, KV per layer

Side by side

ideogram-4-fp8FLUX.1-dev
Parameters9.3B11.9B
Architectureqwen3vlflux
Layers
Native context
Mixture of expertsnono
Quantizations published1937
Smallest quantization4.46 GiB3.75 GiB
Q4_K_M4.68 GiB6.46 GiB
Licenceotherother

KV cache by context

the term that decides long-context viability
Contextideogram-4-fp8FLUX.1-devRatio
4,096
8,192
16,384
32,768
65,536
131,072