Grok 4.7 登顶 Harvey LAB-AA v1.1，压过 GPT-6 Astra

Tesla Owners Silicon Valley @teslaownersSV原推发布于 2026.10.09

在 X 上打开原推

NEWS: Grok 4.7 has claimed the #1 spot on Harvey LAB-AA v1.1. It now sits ahead of Claude Opus 5.5, GPT-6 Astra, Muse Spark 1.3 and other leading models. The challenge is unforgiving: one material hallucination can mean zero points for the entire task. It’s a test of end-to-end accuracy, not just convincing answers. Grok 4.7 is pushing deeper into high-stakes reasoning.