SkillTV-Bench (Trajectory Verification): leaderboard

Metric: Balanced accuracy (%; mean of success-class and failure-class accuracy for the judge's pass/fail verdict over 203 held-out agent trajectories in nine domains). Source: arxiv.org. Saturation forecast: Around March 2027. 3 models tracked.

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)58.6
2Claude Sonnet 4.656
3GPT-5.247.8

Interactive version: theaggregate.ai/benchmark?slug=skilltv-bench-trajectory-verification · How It Works · Data refreshed daily, snapshot 2026-09-26.