19.10.2026

The other award-winning works covers early warning signs of Mediterranean heatwaves and social justice in cycling infrastructure. Each of the recipients was honored with a laudatory speech presented in an engaging format.

Sebastian Willmann was awarded 2nd place for his bachelor thesis, "Leveraging large language models for data analysis in climate research." In his work, he combined large language models (LLMs) with standardized climate data to facilitate rapid and reliable analysis. The chatbot platform he developed, FrevaGPT, offers a user-friendly, dialogue-based interface: scientists can ask questions about climate data and programming code without needing to navigate complex data formats, tools, or programming languages in detail themselves. The chatbot translates intricate research questions into clear, reusable scripts that can be verified and adjusted as needed. In this way, FrevaGPT bridges the capabilities of large language models with the transparency and reliability required in climate research.

Furthermore, Sebastian Willmann has already integrated his system into the DKRZ infrastructure. As part of the "ChatGRT" project, which is funded by the Volkswagen Foundation, researchers from the DKRZ and the GFZ Potsdam are continuing the development of the "FrevaGPT" prototype, now known as "Climate Claw."