That's a very one dimensional phrasing. Symbolics has a ton of things SymPy doesn't have. SymPy has some things that Symbolics doesn't have. Which one is "better" depends on which things you need to use. SymPy still has a few classic math things that Symbolics doesn't (though some aspects like prime number pieces are deliberately left out because it's covered by other existing Julia packages, and IMO shouldn't be a part of SymPy in the first place), and Symbolics has plenty of things that SymPy doesn't have around optimized codegen and parallelized Groebner basis algorithms.
That said, Symbolics was released in 2022. The thread you posted was from 2022. In 2026, it does support limits (https://github.com/SciML/SymbolicLimits.jl), integrals, (https://docs.sciml.ai/SymbolicIntegration/dev/) etc., and in fact that symbolic integration is much more advanced than SymPy, except on symbolic integration of ODEs (surprisingly, though that has been pretty deliberate since that domain almost never has an analytical solution so it's not the most useful to real practitioners and is more of a teaching tool)
That said, Symbolics was released in 2022. The thread you posted was from 2022. In 2026, it does support limits (https://github.com/SciML/SymbolicLimits.jl), integrals, (https://docs.sciml.ai/SymbolicIntegration/dev/) etc., and in fact that symbolic integration is much more advanced than SymPy, except on symbolic integration of ODEs (surprisingly, though that has been pretty deliberate since that domain almost never has an analytical solution so it's not the most useful to real practitioners and is more of a teaching tool)