Accuracy and Completeness of AI Chatbots (ChatGPT-4o vs. Gemini 2.5 Pro) for Leopard Gecko (Eublepharis macularius) Husbandry Information

Large Language Models (LLMs) like OpenAI's ChatGPT-4o and Google's Gemini 2.5 Pro are increasingly used for information retrieval, including incorporation into popular Q&A websites where specialised topics like pet care are often asked. This reliance raises concerns about the accuracy and completeness of AI-generated advice and potential impacts on animal welfare. This study aimed to evaluate and compare the knowledge retrieval capabilities of ChatGPT-4o and Gemini 2.5 Pro concerning husbandry practices for the Leopard Gecko ( Eublepharis macularius ), a popular pet reptile. To our knowledge, this represents the first direct comparison of these specific, recently released AI models for pet reptile husbandry information retrieval. We queried both models with 21 questions on biology, care, and health using two prompting styles: questions asked individually in separate chats and all questions asked within a single chat. Responses were scored for accuracy and completeness. Both models provided generally accurate information (scores ≥ 2/3), with no dangerous inaccuracies observed. However, completeness varied. Gemini 2.5 Pro offered more detailed and nuanced answers. Prompting style drastically affected output: individual questions yielded substantially more comprehensive answers (average words: ChatGPT 136; Gemini 377) than combined prompts (average words: ChatGPT 12; Gemini 64), which resulted in highly concise, often incomplete responses, especially from ChatGPT-4o. The omissions in AI answers highlight risks if users rely solely on concise AI answers. We conclude that while both AI models can be useful starting points, users must critically evaluate the information, understand prompting impacts, and consult expert resources and veterinarians to ensure optimal animal welfare.