Model comparison

MR_MARTIN_31b_V1.0_Gemma-4-base-ftp vs gemma-4-12B-it

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

From the file· summed bytes, KV per layer

Side by side

MR_MARTIN_31b_V1.0_Gemma-4-base-ftpgemma-4-12B-it
Parameters30.7B12.0B
Architecturegemma4gemma4
Layers48
Native context262,144
Mixture of expertsnono
Quantizations published2359
Smallest quantization6.66 GiB3.92 GiB
Q4_K_M6.63 GiB
Licenceotherapache-2.0

KV cache by context

the term that decides long-context viability
ContextMR_MARTIN_31b_V1.0_Gemma-4-base-ftpgemma-4-12B-itRatio
4,0960.72 GiB
8,1920.97 GiB
16,3841.47 GiB
32,7682.47 GiB
65,5364.47 GiB
131,0728.47 GiB