Kim's team stated, "Under the same conditions [as LG AI Research's experiment], Gemini and Grok series models scored approximately 92 points, while ChatGPT and Claude series models scored about 88 ...