Peking University Introduces ContinualSkillBench for AI Agents

Peking University introduced ContinualSkillBench, a new dynamic evaluation framework across five domains to test if AI agents can continually evolve using external skill libraries, revealing that explicit skills do not significantly improve performance.

2026-08-05 ~ 2026-08-06 · 2 related posts

• Peking University Introduces ContinualSkillBench: Evaluating Continual Skill Evolution in LLM Agents — PekingUniversity · 2026-08-05
• ContinualSkillBench: Explicit Skill Libraries Offer Little Advantage for AI Agents — dair_ai · 2026-08-06