Continual-ARC：用 ARC-AGI 三代任务流测试智能体的终身学习能力

流结束后，学习者被冻结，用见过的任务的新实例测记忆、用未见过的新任务测泛化。实验在 Qwen3.5-9B 和 GLM-5.3 Flash 上运行带/不带持久记忆的 agent harness，与两个提示基线、程序归纳法，以及用 Qwen3.5-9B 模仿 GLM-5.3 Flash 的权重训练对比。结果显示，harness 持久记忆在缺乏专家指引时几乎不改善记忆；多数 harness 方法学习效果甚至低于简单缓存每个任务演示的基线。没有 harness 方法在未见任务上获得增益，而权重型学习有显著提升。

Continual-ARC: Measuring Skill Acquisition over a Working Lifetime

Agent harnesses have emerged as a promising mechanism for continual learning, alongside test-time parametric learning. However, the current benchmarks have several limitations: they generally focus on a single applied domain, serve each instance only once, and assume a single family of learner, so memory-based and weight-based systems are rarely compared on the same scale. We propose Continual-ARC, a lifelong stream built from the three ARC-AGI editions, whose tasks need no domain knowledge. The same tasks return after dormant gaps with fresh instances, while new tasks keep arriving. On each instance, the learner may pay for demonstrations, submit answers, or act inside the harness, with retries permitted on incorrect submissions, which are paid for. Each instance is scored based on the share of this feedback budget consumed. Since a learner may keep anything between instances, memory-based and weight-based systems are scored on the same scale. After the stream, the learner is frozen and tested on fresh instances of the tasks it met, which measures memorization, and on unseen tasks, which measures generalization. We run an agent harness with frozen weights, with and without persistence, on Qwen3.5-9B and GLM-5.3 Flash, comparing them with two prompting floors, a program-induction method and Qwen3.5-9B's weights trained to imitate GLM-5.3 Flash. Persistence in the harness generally does not improve memorization unless provided with explicit expert guidance. Most harness-based methods still learn less than a floor that simply caches each task's demonstrations, across all models we tested. No harness-based method gains on unseen tasks, whereas weight-based learning shows substantial gains. This opens the door to future weight-based continual learning that could teach models to drive harnesses in new domains.

This report is categorized under AITOP's "Papers" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Continual-ARC, ARC-AGI, Qwen3.5, GLM.

Source attribution: curated from arXiv cs.LG (@Daniel Lenton). Cross-source corroboration: 2 independent sources have covered this development (including ARC Prize), confirming industry-wide signal strength.