← Back to all builds
Prime Agent
Open-source self-improving coding harness where sub-agents are function calls in a persistent kernel. Scored 95.5% on ARC-AGI-3, above the human expert baseline.
Prime Agent is a general-purpose coding harness On ARC-AGI-3, it scores 95.5%, surpassing the human-expert baseline, but the gain is not benchmark-specific. We see major improvements across models when compared to their proprietary harnesses:
