PagerDuty SRE (Claude Code) - Misdirection: leaderboard
Misdirection (0-1), Claude Code fixed-harness cohort; 17 injected Kubernetes financial-app fault scenarios, one trial per agent, default config except model; same Grafana MELT, source code and alerts, no direct Kubernetes substrate or remediation. Three-provider judge ensemble. Lower is better.
Metric: Misdirection (0-1), Claude Code fixed-harness cohort; 17 injected Kubernetes financial-app fault scenarios, one trial per agent, default config except model; same Grafana MELT, source code and alerts, no direct Kubernetes substrate or remediation. Three-provider judge ensemble. Lower is better.. Source: www.pagerduty.com. Saturation forecast: Around August 2028. 2 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 | 0.56 | #3 |
| 2 | Claude Sonnet 5 | 0.65 | #55 |
No result here: #5 GPT-6 Astra, #7 Gemini 4 Argon, #8 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=pagerduty-sre-claude-code-misdirection · How It Works · Data refreshed daily, snapshot 2026-10-16.