SRE Skills Bench - Kubernetes: leaderboard
SRE Skills Bench — accuracy on Kubernetes operations 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) | 99.3 | #58 |
| 2 | GPT-5.4 | 99.3 | #79 |
| 3 | Claude Opus 4.7 | 99.3 | #49 |
| 4 | GPT-5.3 Codex | 98.6 | #69 |
| 5 | GPT-5.2 Pro | 98.6 | #61 |
| 6 | Kimi K2.5 | 97.9 | #144 |
| 7 | GPT-5.5 (High) | 97.9 | #31 (GPT-5.5) |
| 8 | Claude Opus 4.6 | 97.3 | #63 |
| 9 | Gemini 3 Pro | 97.3 | #80 |
| 10 | Claude Opus 4.5 | 96.6 | #83 |
| 11 | Claude Sonnet 4.6 | 94.5 | #90 |
| 12 | GPT-5.1 Codex Max | 94 | #100 |
| 13 | GPT-5.2 | 93.2 | #110 |
| 14 | GPT-5.1 | 90.2 | #135 |
| 15 | Gemini 3.1 Flash Lite | 89.7 | #191 |
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-kubernetes · How It Works · Data refreshed daily, snapshot 2026-10-20.