New MLS-Bench benchmark introduces 140 tasks across 12 domains to evaluate recursive AI self-improvement
The benchmark addresses a lack of standardized self-evolution testing.
Story Brief
The benchmark addresses a lack of standardized self-evolution testing.

