In　the　era　of　intelligent　interconnection　of　all　things,　society　gradually　forms　a　system　composed　of　numerous　in-telligent　entities　that　collaborate　and　evolve　through　division　of　labor.　However,　in　reality,　high-intelligence　individuals　often　fail　to　achieve　high　system　efficiency　after　interacting　to　form　intelligent　groups.　The　relationship　between　interactions　among　intelligent　individuals　and　the　emergence　of　system　efficiency　remains　to　be　explored　and　answered.　In　this　study,　we　use　color　coordination　game　as　the　basic　scenario　to　investigate　the　impact　of　decision　error　rates　of　different　types　of　intelligent　individuals,　the　proportion　of　the　same　type　of　intelligent　individuals,　and　the　differences　in　decision　error　rates　within　structured　groups　on　problem-solving.　Through　simulation　experiments,　we　find　that　in　high-intelligence　groups,　adding　an　appropriate　number　of　relatively　low-intelligence　individuals　can　optimize　and　stabilize　problem-solving.　Conversely,　adding　lower-intelligence　individuals　to　a　low-intelligence　group　will　decrease　overall　performance.　Notably,　when　the　added　lower-intelligence　indi-viduals　are　at　the　center　of　the　group,　their　impact　on　problem-solving　efficiency　becomes　more　pronounced.　Additionally,　we　discover　that　having　significant　intelligence　differences　within　high-intelligence　groups　and　moderate　intelligence　differences　within　low-intelligence　groups　can　also　enhance　overall　group　efficiency.　This　research　provides　an　explanation　for　the　typical　phenomenon　of　mismatch　between　individual　intelligence　and　system　efficiency　in　collaborative　processes　and　offers　guidance　for　improving　system　efficiency　in　real-world　social　scenarios.　©　2024　IEEE.