Jev 正式发布:面向决策的 System One 模型
TypeSafe 创始人 Diogo Almeida 介绍 Jev、RLCD 训练方法,以及面向软件决策场景的模型定位。原帖附有 2 分 56 秒发布视频。

介绍 Jev-Omni,基于 Gemma 4 12B IT,可实时处理文本、图像、音频和视频,面向 AR 眼镜等空间计算场景。
Text, image, audio, video — judged in real time by one model. Jev-Omni runs on Gemma 4 12B IT and it's built for exactly the instant, multimodal reasoning spatial computing needs. This is the AI layer AR glasses have been missing. #SpatialComputing