Do Agent Optimizers Compound? A Continual-Learning Evaluation on Terminal-Bench 2.0

In 30 Sekunden

Was passiert ist

Researchers tested whether agent optimization methods maintain gains when applied repeatedly to new tasks over time using Terminal-Bench 2.0.

Warum es zählt

Matters for engineers deploying agents in production where continuous optimization happens as new failures and tasks emerge.

Achtung

Only RELAI-VCL maintained and improved gains across optimization rounds; other methods either degraded or plateaued, suggesting most gains may not be stable.