ITBench-AA — leaderboard
Artificial Analysis implementation of IBM's ITBench SRE benchmark, testing AI agents on Kubernetes incident root-cause analysis from offline incident snapshots.
Metric: Average Precision at Full Recall (self-reported). Source: benchmarklist.com. Status: saturation imminent. 23 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 (Max) | 46.7 |
| 2 | GPT-5.5 | 45.8 |
| 3 | Qwen 3.7 Max (Max) | 42.5 |
| 4 | GLM-5.1 | 40.3 |
| 5 | Gemini 3.5 Flash | 40.3 |
| 6 | Claude Sonnet 4.6 (Max) | 39.8 |
| 7 | DeepSeek V4 Pro (Max) | 38.3 |
| 8 | MiMo-V2.5-Pro | 38.2 |
| 9 | Gemma 4 31B | 37.3 |
| 10 | Qwen 3.5 27B | 35.5 |
| 11 | GPT-5.4 Mini | 35.2 |
| 12 | GPT-5.4 | 34.5 |
| 13 | Qwen 3.5 397B A17B | 34.1 |
| 14 | Grok 4.3 | 32.7 |
| 15 | DeepSeek V4 Flash (Max) | 31.5 |
Interactive version: theaggregate.ai/benchmark?slug=itbench-aa · How the rankings work · Data refreshed daily, snapshot 2026-07-22.