这个模型用循环潜在推理做上下文学习，150M 参数在 ARC-AGI-1 上花 0.0007 美元就拿到 29.5% 的 pass@2，性价比直接刷新纪录，值得看看它怎么做到的。
BDH-CQ 是一种结合上下文学习与循环潜在推理的推理模型，推理时输入持续更新模型记忆，并在高维潜在空间迭代计算，不输出中间推理过程。在 ARC-AGI-1 公开评测集上，150M 参数配置达到 29.5% pass@2，单任务推理成本仅 0.0007 美元。该结果突破了 ARC-AGI-1 此前报告的成本-准确率帕累托前沿，树立了新的基准成本效率标杆。研究还通过受控 ARC 干预实验，分析了模型从演示中学习的内容、应用推断变换的一致性及难点概念。
BDH-CQ: In-Context Learning with Recurrent Latent Reasoning
We introduce BDH-CQ, a reasoning model that combines in-context learning with recurrent latent reasoning. Inputs presented at inference time continuously update the model's recurrent memory; the model then solves a query through iterative computation in a high-dimensional latent space, without verbalizing its intermediate reasoning. We evaluate the model on the public ARC-AGI-1 evaluation set and use controlled ARC-like interventions to study what it learns from demonstrations, how consistently it applies an inferred transformation, and which concepts remain difficult. A 150M-parameter configuration reaches 29.5% pass@2 at a computed inference cost of \$0.0007 per task. This operating point breaks through the previously reported ARC-AGI-1 cost-accuracy Pareto frontier, establishing a new state of the art in benchmark cost efficiency.