{"id":"tp-40859","type":"article-newspaper","title":"Google DeepMind’s XYEval Finds Relative AI Performance Falls Up to 46.7%","author":"Simon Yoon","publisher":"TOKENPOST Inc.","url":"https://www.tokenpost.com/news/technology/40859","datePublished":"2026-10-26T07:52:23.000Z","dateModified":"2026-10-26T07:52:23.000Z","language":"en","abstract":"The benchmark tests five models across six task suites. Its largest relative performance drop was for Gemini 3.1 Pro on Terminal-Bench.","section":"Technology","references":[]}