Model comparison

Llama3.2-30B-A3B-II-Dark-Champion-INSTRUCT-Heretic-Abliterated-Uncensored vs gemma-4-12B-it-qat-q4_0-unquantized

These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: gemma-4-12B-it-qat-q4_0-unquantized's KV cache at 32K is 2.4× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

Llama3.2-30B-A3B-II-Dark-Champion-INSTRUCT-Heretic-Abliterated-Uncensoredgemma-4-12B-it-qat-q4_0-unquantized
Parameters30.0B12.0B
Architecturellamagemma4
Layers4748
Native context131,072262,144
Mixture of expertsyes, 8 expertsno
Quantizations published342
Smallest quantization6.07 GiB6.50 GiB
Q4_K_M17.11 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextLlama3.2-30B-A3B-II-Dark-Champion-INSTRUCT-Heretic-Abliterated-Uncensoredgemma-4-12B-it-qat-q4_0-unquantizedRatio
4,0960.73 GiB0.72 GiB1.02×
8,1921.47 GiB0.97 GiB1.52×
16,3842.94 GiB1.47 GiB2.00×
32,7685.88 GiB2.47 GiB2.38×
65,53611.75 GiB4.47 GiB2.63×
131,07223.50 GiB8.47 GiB2.77×