GPT-4V(ision) is a human-aligned evaluation metric for text-to-3D generative models. The method involves two main components: a prompt generator and a 3D assets comparator. The prompt generator, using GPT-4V, creates input prompts tailored to the evaluator's demands, while the 3D assets comparator uses GPT-4V to compare two 3D shapes based on user-defined criteria. These components enable the system to rank text-to-3D models using the Elo rating system. Experimental results show that the proposed metric aligns well with human judgment across various evaluation criteria, providing a versatile and scalable solution for evaluating text-to-3D models. The code for this method is available at <https://github.com/3DTopia/GPTEval3D>.GPT-4V(ision) is a human-aligned evaluation metric for text-to-3D generative models. The method involves two main components: a prompt generator and a 3D assets comparator. The prompt generator, using GPT-4V, creates input prompts tailored to the evaluator's demands, while the 3D assets comparator uses GPT-4V to compare two 3D shapes based on user-defined criteria. These components enable the system to rank text-to-3D models using the Elo rating system. Experimental results show that the proposed metric aligns well with human judgment across various evaluation criteria, providing a versatile and scalable solution for evaluating text-to-3D models. The code for this method is available at <https://github.com/3DTopia/GPTEval3D>.