I have superficially researched both options (with the conclusion that I cannot use either, since my installation would be too small, and would suffer from severe heat loss due to an unfavourable volume-to-surface ratio - it makes sense to design thermal stores for a city or neighbourhood, not a household).
I'd add a few notes:
A thermal store using silicate sand is not limited by the melting point of the sand, but the structural strength of the materials holding the sand. You can count on stainless steel up to approximately 600 C, more if you design with reserve strength and good understanding of thermal expansion/contraction. Definitely don't count on anything above 1000 C or forget the word "cheap". I have read about some folks designing a super-hot thermal store, but they plan to heat graphite (self-supporting solid material) in an inert gas environment.
Heat loss intensifies with higher temperatures, and the primary type of heat loss becomes radiative loss. Basically, stuff starts glowing. For example, the thermal conductivity of stone wool can be 0.04 W / mK at 10 C, and 0.18 W / mK at 600 C.
Water can be kept liquid beyond 100 C. The most recent thermal stores in Finland are about 100 meters below surface, where the pressure of the liquid column allows heating water to 140 C.
However, any plan of co-generation (making some electricity while extracting the stored heat) requires solid materials and high temperatures.