GPT-4.5 Preview takes first place in the Elimination Game Benchmark, which tests social reasoning (forming alliances, deception, appearing non-threatening, and persuading the jury).
GPT-4.5 Preview takes first place in the Elimination Game Benchmark, which tests social reasoning (forming alliances, deception, appearing non-threatening, and persuading the jury).
![[link]](https://i.redd.it/44ms4kfr2ime1.png){kind=link}