o11y-bench - Pass^3: leaderboard
Grafana Labs' o11y-bench runs agents on 63 observability and SRE tasks, three attempts per task, with each model and reasoning setting listed separately. Pass^3 is the percentage of tasks passed on all three attempts; higher is better.
Metric: Tasks passed on all three attempts, Pass^3 (%). Source: o11ybench.ai. Saturation forecast: Around February 2027. 71 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 (Medium) | 93.75 | #3 (Claude Opus 5.5) |
| 2 | Claude Sonnet 5.5 (xHigh) | 90.62 | #14 (Claude Sonnet 5.5) |
| 3 | Claude Sonnet 5.5 (High) | 89.06 | #14 (Claude Sonnet 5.5) |
| 4 | Claude Opus 5.5 (Max) | 87.5 | #3 (Claude Opus 5.5) |
| 5 | GPT-6 Astra (Low) | 87.5 | #6 (GPT-6 Astra) |
| 6 | GPT-6.1 Sol (Medium) | 87.5 | #7 (GPT-6.1 Sol) |
| 7 | Claude Haiku 5.5 (xHigh) | 87.5 | #43 (Claude Haiku 5.5) |
| 8 | Gemini 4 Argon (Low) | 87.5 | #9 (Gemini 4 Argon) |
| 9 | GPT-6 Astra (Medium) | 85.94 | #6 (GPT-6 Astra) |
| 10 | GPT-6.1 Sol (xHigh) | 85.94 | #7 (GPT-6.1 Sol) |
| 11 | GPT-6.1 Sol (High) | 85.94 | #7 (GPT-6.1 Sol) |
| 12 | Claude Haiku 5.5 (High) | 85.94 | #43 (Claude Haiku 5.5) |
| 13 | Claude Opus 5.5 (Low) | 85.94 | #3 (Claude Opus 5.5) |
| 14 | GPT-6 Astra (xHigh) | 84.38 | #6 (GPT-6 Astra) |
| 15 | Claude Sonnet 5.5 (Max) | 84.38 | #14 (Claude Sonnet 5.5) |
No result here: #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=o11y-bench-pass-pow-3 · How It Works · Data refreshed daily, snapshot 2026-11-05.