Google DeepMind’s Gemini Earns Gold at the 2025 International Mathematical Olympiad

[Redaktion]

Google DeepMind's Gemini solved five of six IMO 2025 problems, earning gold-level recognition. It produced solutions in natural language within the allotted 4.5-hour contest window.

https://www.notebookcheck.net/Google-DeepMind-s-Gemini-Earns-Gold-at-the-2025-International-Mathematical-Olympiad.1064843.0.html

[Worgarthe]

And yet it still cannot beat a semi-competent chess player in a hyper bullet or bullet format. No even need to mention how it gets utterly destroyed by anyone with 2200+ ELO. Fun stuff.