Aletheia aborda o FirstProof de forma autónoma.
Aletheia tackles FirstProof autonomously
February 24, 2026
Autores: Tony Feng, Junehyuk Jung, Sang-hyun Kim, Carlo Pagano, Sergei Gukov, Chiang-Chiang Tsai, David Woodruff, Adel Javanmard, Aryan Mokhtari, Dawsen Hwang, Yuri Chervonyi, Jonathan N. Lee, Garrett Bingham, Trieu H. Trinh, Vahab Mirrokni, Quoc V. Le, Thang Luong
cs.AI
Resumo
Relatamos o desempenho do Aletheia (Feng et al., 2026b), um agente de pesquisa matemática baseado no Gemini 3 Deep Think, no desafio inaugural FirstProof. Dentro do prazo estipulado pelo desafio, o Aletheia resolveu autonomamente 6 problemas (2, 5, 7, 8, 9, 10) de um total de 10, de acordo com a avaliação da maioria dos especialistas; observamos que os especialistas não foram unânimes apenas em relação ao Problema 8. Para total transparência, explicamos nossa interpretação do FirstProof e divulgamos detalhes sobre nossos experimentos e nossa avaliação. Os *prompts* e resultados brutos estão disponíveis em https://github.com/google-deepmind/superhuman/tree/main/aletheia.
English
We report the performance of Aletheia (Feng et al., 2026b), a mathematics research agent powered by Gemini 3 Deep Think, on the inaugural FirstProof challenge. Within the allowed timeframe of the challenge, Aletheia autonomously solved 6 problems (2, 5, 7, 8, 9, 10) out of 10 according to majority expert assessments; we note that experts were not unanimous on Problem 8 (only). For full transparency, we explain our interpretation of FirstProof and disclose details about our experiments as well as our evaluation. Raw prompts and outputs are available at https://github.com/google-deepmind/superhuman/tree/main/aletheia.