GPT-6.1 Sol Locates Brain Landmarks in Three Live Mouse Procedures
Exin Therapeutics reported that OpenAI's GPT-6.1 Sol participated in three consecutive live-mouse brain procedures, using surgeon feedback and prior images, measurements, and corrections to improve anatomical-landmark localization. The team said identification time fell from 169 to 86 seconds and localization error dropped from 0.6 to 0.3 millimeters by the third procedure, while the number of manual corrections fell from four to two. Licensed veterinarians retained control, approved irreversible steps, and supervised every operation; the researchers said three procedures were insufficient to establish stable performance.
This is a concrete early test of GPT-6.1 Sol assisting with real surgical equipment under strict human oversight, but the tiny sample, vendor-reported results, and lack of independent validation or peer review limit the evidence.
北京时间 11 月 8 日，据《WIRED》此前报道及生物科技公司 Exin Therapeutics 最新披露，OpenAI 旗下 AI 模型 GPT-6.1 Sol 已被用于活体小鼠脑部手术实验，并在连续三次操作中，通过外科医生的反馈改善了定位表现。到第三次操作时，其识别关键解剖标志的耗时从 169 秒缩短至 86 秒，定位误差从 0.6 毫米降至 0.3 毫米。
Exin Therapeutics 联合创始人兼 CEO 加布里埃尔 · 奥卡尼亚-桑特罗 (Gabriel Ocana-Santero)11 月 5 日在社交平台 X 上公布了这一结果。研究团队让 GPT-6.1 Sol 连续参与三次活体小鼠脑部手术中的解剖标志定位任务。每次操作结束后，专业外科医生都会纠正 AI 识别的位置，模型则保存相关图像、测量数据及修正记录，供下一次操作参考。