Agent Harnesses for Continual Learning
Agent harnesses have emerged as a promising mechanism for continual learning, alongside test-time parametric learning.
Limitations of Current Benchmarks
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.
Introducing Continual-ARC
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.
Interaction and Scoring
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.
Post-Stream Evaluation
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.
Experimental Setup
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.
Findings and Conclusion
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.