VerifAI's open-source MultiLLM framework generates Code by calling LLMs in parallel and ranks their outputs to find the best results. MultiLLM can be extended to support new LLMs and custom ranking function to evaluate a variety of outputs from LLMs.