跳到正文
稳定报道时间 2026-10-02 12:00

Kepler: 可审计的世界模型用于 ARC-AGI-3

Kepler 是一个开源工具,它使用可执行的世界模型来评估交互环境中的智能体,在一组公共游戏中实现了高性能。

01

证据

  • AarXiv cs.AI一手来源2026-10-02 12:00
    Kepler obtained a server-verified 100.00 RHAE on all 25 public games
    查看来源
  • AarXiv cs.AI一手来源2026-10-02 12:00
    Abstract: ARC-AGI-3 evaluates agents in interactive environments whose rules and objectives must be inferred from observation. We present Kepler, an open-source harness that represents hypotheses as executable world models and validates them through retrospective transition checks and conditional prediction checks. Under one frozen Claude Opus 5 configuration, Kepler obtained a server-verified 10…
    查看来源