OpenAI GPT-6.1 Sol halves mouse brain-surgery positioning error after three operations
Briefing by News4U · First published
ITHome reports that OpenAI’s GPT-6.1 Sol participated in three consecutive live-mouse brain surgeries, using surgeon feedback to cut positioning error from 0.6 mm to 0.3 mm and halve the number of required corrections. The model also reduced the time to identify key anatomical landmarks from 169 seconds to 86 seconds and lowered the number of manual corrections from four to two. Licensed veterinarians still performed incisions, craniotomies and suturing, while the AI only handled landmark localization and injection planning under human supervision. [1]
Sources
- OpenAI GPT-6.1 Sol 参与活体小鼠脑手术：连续操作三次，定位误差减半
ITHome · · Summary only