Can OpenSource beat ChatGPT? -- A Comparative Study of Large Language Models for Text-to-Code Generation