Maxim Lott's analysis, when Claude 3 was released, is the basis for this question. A similar analysis will likely be published when GPT-5 is released, and that will be the basis for resolving this prediction.
https://www.maximumtruth.org/p/ais-ranked-by-iq-ai-passes-100-iq
🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ101 | |
| 2 | Ṁ76 | |
| 3 | Ṁ74 | |
| 4 | Ṁ28 | |
| 5 | Ṁ27 |
Alright, based on trackingai.org (Maxim's site for analyzing and aggregating model IQ estimates), GPT-5 Thinking scored a 130 on the Mensa Norway test and 116 on their own offline test question set (Aug 13, 2025)
Resolving this to yes.
@PhilosophyBear my question was based on this article from Maxim Lott (https://www.maximumtruth.org/p/ais-ranked-by-iq-ai-passes-100-iq)
If you have a measurement instrument in mind, please suggest -- otherwise, I would use a similar analysis to this once it is released (which will surely be done for GPT-5).