What will be the best performance on FrontierMath Tier 4 by December 31st 2025?
10
Ṁ990
Dec 31
18%
0% - 10%
38%
10 - 20%
27%
20 - 30%
6%
30 - 40%
3%
40 - 50%
1.8%
50 - 60%
1.8%
60 - 70%
1.8%
70 - 80%
1.8%
80 - 90%
1.8%
90 - 100%

The best performance by an AI system on FrontierMath Tier 4 as of December 31st 2025. See https://epoch.ai/frontiermath, under the section Tier 4, for results accepted for the purpose of this market. The "performance" is measured in terms of Pass@1 Accuracy.

At market creation (and day of the official announcement of the benchmark), the best model is o4-mini (high), with a score of 6.25%.

See also best performance on FrontierMath Tier 1-3:

AITechnical AI TimelinesAI BenchmarksAI 2025 Forecasting Survey by AI Digest
Get Ṁ1,000 play money
4 Comments
Sort by:
bought Ṁ50 10 - 20% YES

GPT-5 (high) got one more problem than previous models, for a score of 8.3% (4/48).

My intuition is that you can't get much better results on this benchmark just by scaling current methods and that no one will implement a new method before end of 2025.

I think there's still a fair bit it can be pushed but I agree probably no one will implement a new method before end of 2025. do you think the method that first gets 30% will use a new method?

@Bayesian Yeah. But I didn't do too much research on the questions I just know they are unique from trainable datasets, and require a lot of reasoning steps. I think we need a new method that will help AIs better generalize their learnings and skills from different domains.

Related questions