Why are the SOTA open-weight models scoring (relatively) low scores on AA-Omniscience Index

The Reddit community is discussing why state-of-the-art (SOTA) open-weight models are achieving relatively low scores on the AA-Omniscience Index. Users are surprised by these scores, noting they are significantly lower than those of Gemini-3.* flash models. This observation has sparked conversation regarding the performance metrics and comparative evaluations within the AI development community.