OpenRCA: leaderboard
OpenRCA RCA-Agent root-cause analysis accuracy on software failures using logs and telemetry. Balanced prompting and oracle-selected telemetry are separate evaluation protocols.
Metric: Accuracy (%). Source: microsoft.github.io. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 36.42 | #63 |
| 2 | Claude Opus 4.5 | 28.36 | #83 |
| 3 | GPT-5.2 | 19.4 | #110 |
| 4 | Gemini 3 Pro | 12.54 | #80 |
| 5 | Claude 3.5 Sonnet | 11.34 | #350 |
| 6 | GPT-4o | 8.96 | #343 |
| 7 | Gemini 1.5 Pro | 2.69 | #445 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=openrca · How It Works · Data refreshed daily, snapshot 2026-10-20.