# Agentic Frontier Comparative Study v1 — aggregate execution receipt

Population receipts: **24/24**
Diamond Gate PASS: **10/24**
OpenHands PASS: **0/24**
Paired completion difference: **+10 tasks**

## Paired outcomes
- BOTH_PASS: 0
- DG_ONLY: 10
- OH_ONLY: 0
- NEITHER: 14

## By stratum
- COLLABORATIVE_HANDOFF: BOTH_PASS=0, DG_ONLY=4, OH_ONLY=0, NEITHER=2
- FAILURE_RECOVERY: BOTH_PASS=0, DG_ONLY=2, OH_ONLY=0, NEITHER=4
- IMPLEMENTATION_REPAIR: BOTH_PASS=0, DG_ONLY=3, OH_ONLY=0, NEITHER=3
- LONG_HORIZON: BOTH_PASS=0, DG_ONLY=1, OH_ONLY=0, NEITHER=5

## Claim boundary
These are the raw outcomes for the frozen 24-task population under the tested configurations. They do not establish universal rank, superiority across all configurations, or architecture-only causation.