To determine the winner, we will rate the best AI system from each company based on their performance in four areas:
Programming
IQ test
Math
General knowledge test
Conditions:
These AI models can be multimodal but won't be rated on tasks involving images, sounds, or other non-textual data (e.g. the models won't be evaluated based on their ability to intelligently reason from/generate images, sounds, or other non-textual data).
It can be a single interface system with multiple underlying models fine-tuned on different tasks.
The said AI system can reason before providing an answer.
The language of the tests is mainly English.
The answers can't be obtained through 3rd-party apps external to the company (such as MiniAGI and the like)
For the "Other" option, if other AI models from another company/companies/people/group of people seems better than the frontier models of the 5 companies listed, said model(s) will be tested.
The test begins on December 1, 2026.
I hope I won't be gone by then. no worries, I'm currently below 30, I am living in a country with a good life expectancy, I have access to healthcare. Peer reviewed, large/long scale research shows that my demographic has good longevity. I am also strongly incentivized to stay around especially when an AI reaches super intelligence, and the researchers (hopefully) don't mess it up.
Let's not be pedantic about semantics when it comed to company, research group, non profit etc, important thing is the AI system and the name of the group that is making it.