Will an AI model achieve superhuman ELO on Codeforces by the 31 December 2025?

Ṁ100Ṁ3.2k

Dec 31

35%

chance

ALL

This market predicts whether any AI model will surpass the current best human competitor's ELO rating on Codeforces by the end of 2025. As of 31 December 2024, the current highest human ELO is 3985 and the current highest AI ELO is 2727 which was achieved by OpenAI's o3 model.

Resolution Criteria: Market resolves YES if any AI lab announces and verifies that their model has achieved a Codeforces ELO rating greater than 3985 by December 31, 2025. Otherwise, resolves NO.

References:

Market context

Technology

Technical AI Timelines

OpenAI

IMO Grand Challenge

Get

1,000

to start trading!

People are also trading

AI model achieves superhuman ELO on Codeforces by June 1st 2027?

87% chance

Once a model achieves superhuman performance on a competitive programming benchmark, will it be less than 5 years before there are "entry level" AI programmers in industry use?

92% chance

Will an AI achieve >85% performance on the FrontierMath benchmark before 2027?

35% chance

Will a large language model beat a super grandmaster at chess by EOY 2035?

70% chance

Will an AI achieve >80% performance on the FrontierMath benchmark before 2027?

48% chance

Will an AI by OpenAI beat a super grandmaster playing chess by 2028?

57% chance

Will a large language model beat a super grandmaster playing chess by EOY 2028?

60% chance

Benchmark Gap #3: Once a model achieves superhuman performance on a competitive programming benchmark, will it be less than 2 years before there are "entry level" AI programmers in industry use?

73% chance

Will an AI achieve >85% performance on the FrontierMath benchmark before 2028?

60% chance

Will an AI model be capable of superhuman persuasion before 2034.

Sort by:

Is anyone actively trying to do this? It probably can, but also probably nobody has. How to resolve @traders

@Gen https://bab407.com.au/blog/post

There is no update on this since 2024, from my humble search using Perplexity I couldn’t find and proof for yes.

@Gen Some discussion here, also no proof of yes. https://manifold.markets/Bayesian/top-score-on-codeforces-by-an-ai-mo#kz58bq4q28

bought Ṁ15 YES

Might be too soon but let’s give it a try

I have a simple question: is CodeForces able to detect when a human is using AI assistance? If not, then this will never happen, because if some AI becomes superhuman, then some humans will definitely use it to boost their own scores.