With the announcement of the 2025 Pritzker Prize winner less than 24 hours away, Archinect has continued an experiment we began last year, asking four artificial intelligence models to predict the 2025 winner. Below, ChatGPT, Claude, Gemini, and Perplexity have their say after (human) Archinect readers offered us their predictions last week.Beyond the novelty of the experiment and curiosity over how accurate the results may turn out to be, eliciting predictions from AI models offers an opportunity to track the capabilities of such systems over time, including accuracy.In 2024, the AI models we consulted struggled with providing a list free of previous winners or, in Gemini's case, not generating the requested number of potential names. This time, ChatGPT continued to offer previous winners, only generating a valid list on the third attempt. Gemini and Claude also provided previous winners on their first attempt. After a brief back-and-forth, however, all models returned ten poten...