SRE Skills Bench - IAM: leaderboard
SRE Skills Bench — accuracy on AWS IAM identity-and-access tasks.
Metric: Accuracy (%). Source: sreskillsbench.com. Saturation forecast: Estimated already saturated. 23 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 100 | #58 |
| 2 | GPT-5.4 | 98.4 | #79 |
| 3 | GPT-5.3 Codex | 98.4 | #69 |
| 4 | Claude Opus 4.7 | 97.7 | #49 |
| 5 | GPT-5.5 (High) | 97.7 | #31 (GPT-5.5) |
| 6 | Kimi K2.5 | 96.9 | #144 |
| 7 | Gemini 3 Pro | 96.9 | #80 |
| 8 | GPT-5.2 Pro | 94.5 | #61 |
| 9 | Claude Opus 4.6 | 92.2 | #63 |
| 10 | GPT-5.2 | 91.4 | #110 |
| 11 | GPT-5.1 Codex Max | 91 | #100 |
| 12 | Claude Opus 4.5 | 90.6 | #83 |
| 13 | GPT-5.1 | 88.6 | #135 |
| 14 | Claude Sonnet 4.6 | 86.9 | #90 |
| 15 | Qwen 3 VL 235B A22B | 86.5 | #238 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=sre-skills-bench-iam · How It Works · Data refreshed daily, snapshot 2026-10-20.