Turns out GPT-5.6 Sol is actually SoTA on ARC-AGI-3. Just took two setting changes. You just have to allow it to reason and work over multiple context windows with the help of our canonical compaction implementation….

OpenAI 通过两个设置调整——允许模型跨多个上下文窗口推理并配合一种规范压缩方法——使 GPT-5.6 Sol 在 ARC-AGI-3 基准上达到了最先进水平(SoTA)。这项发现并非通过重新训练模型实现,而是在推理阶段优化了策略,揭示了大模型推理能力的潜在边界。








