Which of the ten research problems were solved under the organisers’ criteria?
Model / AI: Multiple systems, Aletheia
Model / AI: Multiple systems
organizers judged two of ten problems definitively solved, and one of those duplicated an existing archived proof, against labs claiming five or six on the same ten questions; their summary of the submissions was "a lot of very convincing nonsense".
Source review pending. Imported from the original notes; linked claims and artifacts have not been re-audited in this restructuring.
Model / AI: Aletheia
Google DeepMind's agent solved 6 of the 10 inaugural First Proof problems fully autonomously by majority expert assessment; the graders were academic mathematicians the team recruited, some Google-affiliated, best-of-2 selection used the team's own expertise, and one agent's attempt was rated "Critically Flawed" on a problem the paper counts as solved.
Source review pending. Imported from the original notes; linked claims and artifacts have not been re-audited in this restructuring.