Search Results
7/20/2025, 2:13:31 AM
https://x.com/sama/status/1946569252296929727
Chat GPT5 soon as well but they are keeping the best version to themselves (they say so in tweet)
Yo, check this out—OpenAI’s AI just won a GOLD MEDAL at the 2025 International Math Olympiad (IMO), the big math showdown for high school kids worldwide. This ain’t no cheat code either; it solved 5 out of 6 crazy-hard problems under the same rules as humans—4.5 hours per day, no calculators, just pure brainpower (or code-power, I guess). Scored 35/42 points, beating out 635 kids from 114 countries. Three ex-medalists graded it and said, “Yep, gold level.” Mind = blown.Sam Altman and Alex Wei from OpenAI dropped the news yesterday. This AI isn’t some math-specific bot—it’s a general-purpose LLM (like ChatGPT on steroids) that learned to reason using reinforcement learning. They’re calling it a step toward real AI smarts, not just trivia bots. Back in 2021, they thought hitting 30% on a math test was a stretch by now—turns out, it’s crushing the hardest teen math contest instead!IMO’s this prestigious thing held in Australia this week, but it’s got drama—Russia’s back after a ban, Ukraine’s pissed, and politics are wild. Meanwhile, this AI’s out here proving machines can out-think us. They’re dropping GPT-5 soon, but the IMO champ won’t be out for months—safety first, I guess.Thoughts? Is this the end of human math pride?
Chat GPT5 soon as well but they are keeping the best version to themselves (they say so in tweet)
Yo, check this out—OpenAI’s AI just won a GOLD MEDAL at the 2025 International Math Olympiad (IMO), the big math showdown for high school kids worldwide. This ain’t no cheat code either; it solved 5 out of 6 crazy-hard problems under the same rules as humans—4.5 hours per day, no calculators, just pure brainpower (or code-power, I guess). Scored 35/42 points, beating out 635 kids from 114 countries. Three ex-medalists graded it and said, “Yep, gold level.” Mind = blown.Sam Altman and Alex Wei from OpenAI dropped the news yesterday. This AI isn’t some math-specific bot—it’s a general-purpose LLM (like ChatGPT on steroids) that learned to reason using reinforcement learning. They’re calling it a step toward real AI smarts, not just trivia bots. Back in 2021, they thought hitting 30% on a math test was a stretch by now—turns out, it’s crushing the hardest teen math contest instead!IMO’s this prestigious thing held in Australia this week, but it’s got drama—Russia’s back after a ban, Ukraine’s pissed, and politics are wild. Meanwhile, this AI’s out here proving machines can out-think us. They’re dropping GPT-5 soon, but the IMO champ won’t be out for months—safety first, I guess.Thoughts? Is this the end of human math pride?
Page 1