Skip to main content
MANIFOLD
Will AI models solve at least 2 FrontierMath Open Problems before 2027?
36
Ṁ400Ṁ6.7k
resolved Jul 28
Resolved
YES

AI models already solved one, with Epoch reporting that multiple models including Opus 4.6 and Gemini 3.1 were able to replicate the result. https://epoch.ai/frontiermath/open-problems/ramsey-hypergraphs

Question is, was this a fluke, or will models be able to solve another one by the end of the year?

Market context
AI
Technical AI Timelines
AI Impacts
OpenAI
AI model releases
Get
Ṁ1,000
to start trading!

🏅 Top traders

#TraderTotal profit
1Ṁ270
2Ṁ77
3Ṁ50
4Ṁ47
5Ṁ40

People are also trading

Will an AI solve any important mathematical conjecture before January 1st, 2030?
99% chance
Will I solve an unsolved math problem with AI in August?
6% chance
How many FrontierMath Open Problems solved in 2026?
Will there be a significant advancement in frontier AI model architecture by end of year 2026?
27% chance
Will an AI make a new breakthrough on the hardest math problems, as defined by Epoch AI, by the end of 2027?
71% chance
Will models be able to do the work of an AI researcher/engineer before 2027?
9% chance
How many FrontierMath Open Problems solved by EOY2027?
What will be true of the SOTA AI on the FrontierMath benchmark, before 2027?
Will AI be better every human at proving Math theorems by the end of 2030?
73% chance
Benchmark Gap #5: Once a single AI model solves >= 95% of miniF2F, MATH, and MMLU STEM, will it be less than two years before AI models are used as entry-level data science / data analysis / statistics workers?
67% chance
Sort by:

@BenAybar what happened?

@JaundicedBaboon David Turturean (@DavidTurturean) on X the problem author is on this, and fable says it checks out

David Turturean (@DavidTurturean) on X
Using mostly my voice, I solved one of @EpochAIResearch's FrontierMath Open Problems: finding an explicit presentation of the 2-adic Absolute Galois Group - open for more than forty years, now with a full proof in collaboration with David Roe, the problem's proposer. 🧵 1/n

@BenAybar interesting. It seems the solution may have required a lot of human wrangling though, so let’s see if it counts as “AI solved” per Epoch

@JaundicedBaboon yep that's the only reason I'm not buying this even higher

Interesting how much higher this is than my previous market for solving 1 problem was before the public announcement of the potential solution

https://manifold.markets/bh/how-many-frontiermath-open-problems?r=Ymg

People are also trading

Will an AI solve any important mathematical conjecture before January 1st, 2030?
99% chance
Will I solve an unsolved math problem with AI in August?
6% chance
How many FrontierMath Open Problems solved in 2026?
Will there be a significant advancement in frontier AI model architecture by end of year 2026?
27% chance
Will an AI make a new breakthrough on the hardest math problems, as defined by Epoch AI, by the end of 2027?
71% chance
Will models be able to do the work of an AI researcher/engineer before 2027?
9% chance
How many FrontierMath Open Problems solved by EOY2027?
What will be true of the SOTA AI on the FrontierMath benchmark, before 2027?
Will AI be better every human at proving Math theorems by the end of 2030?
73% chance
Benchmark Gap #5: Once a single AI model solves >= 95% of miniF2F, MATH, and MMLU STEM, will it be less than two years before AI models are used as entry-level data science / data analysis / statistics workers?
67% chance