jevIMAGE
Typesafe 官方晒 Jev 在重复问题匹配等任务上的成绩
官方称 Jev 以远低价格超过前沿模型：重复问题匹配 70%→97%，费用分类 50%→86%，升级判断误报更少。
TypeSafe AI
@typesafeai2026.09.26收录原推发布于 2026.09.25
TypeSafe AI @typesafeai原推发布于 2026.09.25
Jev generally out-performed frontier LLMs at a fraction of the price: Repeat-question matching: 70% → 97% Expense categorization: 50% → 86% (vs human reviewers) Escalation: same catches, fewer false alarms
x.com/typesafeai/status/2103321894661595551采集于 2026.09.26 10:13