跳到正文
@testingcatalog· @testingcatalog · X·· 22 天前AI 评分50
AI 导读

Apodex 发布 Apodex 1.1 模型系列,面向长程任务,将推理从研究报告带入实际任务执行。该版本可处理上传文件、搜索和代码环境,并能生成每条结论都可回溯来源的文件;其 Agent Team 配置在 APEX-Agents、FrontierScience-Research 和 BioMysteryBench 上的得分超过 Apodex 1.0 两倍以上。用户可通过 https://t.co/5BE701uMNI 在线试用。

正文

Apodex has released Apodex 1.1, a model family for long-horizon work that moves reasoning out of research reports and into real task execution.

> Apodex 1.1 can work with uploaded files, search, and code environments.

> It can produce files where every claim is traced back to its source.

> The Agent Team configuration more than doubles the Apodex 1.0 score on APEX-Agents, FrontierScience-Research, and BioMysteryBench.

Try Apodex 1.1 online
https://t.co/5BE701uMNI

来源:@testingcatalog · x.com