In his article »Revealing the Hidden Economics of Open Models in the AI Era« Frank Nagle summarizes the study »The Latent Role of Open Models in the AI Economy«, which he co-authored with Daniel Yue from Georgia Tech.

He shows how clearly open AI models are underestimated in the current market. The basis is an unusually extensive dataset on usage, pricing, and performance of various models.

The authors make it clear that closed models do shape the market, but not primarily due to technical superiority. The decisive factors are instead habits, risk aversion, and a lack of information. At the same time, open models are quickly catching up in terms of performance, are significantly cheaper, and could enable annual savings in the double-digit billions for the global AI economy.

This contrast between perceived and actual value creation runs through the entire text. It points to parallels with early open-source development and indicates what still largely unused significance open models could assume in the AI field. The summary therefore provides a clear reason to examine the underlying study more closely.

The text was automatically translated from German into English. The German quotations were also translated in sense.