CrestingRadarGet featured →

Rank #559 · on radar since 2026-07-03

RewardHarness

Self-evolving agentic reward framework for image-editing evaluation — 47.4% on EditReward-Bench from only 100 preference demos, no reward-model training. arXiv 2605.08703.

Visit homepage ↗preference-learningreward-modelself-evolvingqwen-vl+5GitHub

Momentum

42.2
24h7d

Why it's ranked

Every score decomposes into published factors — the same math for every tool, paid or not. Read the methodology →

Velocity (weighted, cohort-normalized)0.438
Signal decay0.995
Corroboration1.000
Quality gate1.000

Raw signals (30 days)

github · forks1 latest · 2 snapshots
github · stars14 latest · 2 snapshots