Grok 4.7 在 ARC-AGI-3 得分 1.8%
@SpaceXAI On ARC-AGI-3, Grok 4.7 scores 1.8% (vs Grok 4.7's 2.1%) in the standard harness, which lets models carry forward notes between turns, and 10.0% in a n
AI 导读
@SpaceXAI 在 ARC-AGI-3 上，Grok 4.7 在标准 harness 中得分 1.8%（Grok 4.7 为 2.1%），该 harness 允许模型在各轮之间保留笔记；在新的 provider adapter harness 中得分 10.0%，该 harness 保留不透明推理并启用自动压缩。
正文
当前显示原文，中文全文译文尚未就绪。
@SpaceXAI On ARC-AGI-3, Grok 4.7 scores 1.8% (vs Grok 4.7's 2.1%) in the standard harness, which lets models carry forward notes between turns, and 10.0% in a new provider adapter harness, which preserves opaque reasoning and enables auto compaction.