Ballon, M., Algaba, A., Ginis, V. (2026). The relationship between reasoning and performance in large language models—o3 (mini) thinks harder, not longer. Scientific Reports. https://doi.org/10.1038/s41598-026-50923-2