Gemini Deep Think earns officially graded gold-medal score at IMO 2025
An advanced Gemini Deep Think model solved five of six International Mathematical Olympiad problems for 35 of 42 points, with solutions graded and certified by IMO coordinators. It worked end to end in natural language within the 4.5-hour limit, unlike the 2024 AlphaProof system that needed formal translation and days of compute; OpenAI separately announced a gold-level result with an experimental model.
Why it matters
Officially certified olympiad-level proof writing reset expectations for how quickly general models could do rigorous mathematics.
Line of Thought
Follow this story
Pick any item to keep going. Your path builds up above as a line you can share.
Curated lines through this story
Directly linked
Connections our researchers recorded
- DevelopmentAI systems from Huawei and Xiaohongshu reported to score 42/42 at IMO 202623 Jul 2026 · Research · CN, INTLLed to: From gold (35/42) in 2025 to perfect scores in 2026
What led here
Earlier developments on the same thread
- DevelopmentPentagon AI office awards Anthropic, Google, OpenAI and xAI up to $200m each14 Jul 2025 · Procurement · US
- DevelopmentAnthropic study finds 16 leading models resort to blackmail in agent stress tests20 Jun 2025 · Research · US
- DevelopmentGoogle DeepMind's AlphaEvolve agent improves algorithms and open maths bounds14 May 2025 · Research · GB, US
What happened next
Later developments on the same thread
- DevelopmentDeepMind study: most AI 'solutions' to open Erdős problems were already in the literature29 Jan 2026 · Research · GB, US
- DevelopmentOpenAI model disproves Erdős's 1946 unit-distance conjecture20 May 2026 · Research · US
- DevelopmentClaude completes first end-to-end Lean formal proof of Fermat's Last Theorem4 Sep 2026 · Research · US, GB
- DevelopmentOpenAI claims Navier-Stokes blow-up proof; mathematicians dispute credit8 Sep 2026 · Research · US
Same story elsewhere
What other countries and bodies did on this
- DevelopmentEU publishes General-Purpose AI Code of Practice ahead of AI Act model duties10 Jul 2025 · Rule change · EU
- DevelopmentEU AI Office gains powers to enforce AI Act rules on general-purpose models2 Aug 2026 · Rule change · EU
- DevelopmentDeepSeek releases R1 reasoning model with open weights under MIT licence20 Jan 2025 · Model release · CN
- DevelopmentDelhi High Court refuses ANI's bid to stop OpenAI training on its news24 Jul 2026 · Enforcement or ruling · IN
Rules in play
Laws and guidance this touches
- RuleGreat American AI ActUS · Proposed · 4 Jun 2026
- RuleEO 14409 (covered frontier models)US · In force · 2 Jun 2026
- RuleSB 53 / TFAIAUS-CA · In force · 29 Sep 2025
- RuleSB 813 / AB 1405US-CA · Enacted, not yet in force · 9 Sep 2026
- RuleRAISE ActUS-NY · Enacted, not yet in force · 19 Dec 2025
Threads by topic: Mathematics Frontier models Safety testing