Detection Made Easy: Potentials of Large Language Models for Solidity Vulnerabilities