FaulT-Bench - Wrong-Cause Tickets - Outcome Score: leaderboard
FaulT-Bench (arXiv 2608.27021v1): network troubleshooting agents diagnose emulated networks from trouble tickets that point at the wrong cause, in 24 scenarios over three passes, judged by gpt-5-mini; the class had no timed-out runs. Outcome score (Table 4): 0.7 diagnosis accuracy plus 0.3 explanation quality, 0-1. Higher is better.
Metric: Judged outcome score (0-1; 0.7 diagnosis accuracy + 0.3 explanation, answered runs, 0% timeouts). Source: arxiv.org. Saturation forecast: Around February 2027. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 0.98 | #94 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=fault-bench-wrong-cause-tickets-outcome-score · How It Works · Data refreshed daily, snapshot 2026-11-05.