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