Terminal X's Eval Agent outperforms GPT-6.1 in accuracy of investment decisions, matching senior investor judgment by 88.6%. This improvement is attributed to the use of a 54-criterion rubric for evaluating output quality.
מקור: OpenAI News — לכתבה המלאה
כתיבת תגובה