Power Systems Agent Benchmark (Aider): leaderboard
Power Systems Agent Benchmark (arXiv 2606.20950, 2026) has a tool-using agent solve 41 electric power engineering task families, from power flow and protection to dispatch and restoration, by submitting structured decisions that a deterministic evaluator recomputes. Each case scores against its reference optimum (infeasible scores zero); this board is the mean on a 0-100 scale for one run per model under the Aider agent on the public cases, higher is better.
Metric: Mean optimum-normalized score (0-100). Source: github.com. Saturation forecast: Around February 2027. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2.5 | 85.12 | #252 |
| 2 | Qwen3 Coder Next | 53.42 | #711 |
| 3 | Claude Haiku 4.5 | 51.22 | #568 |
| 4 | GLM-4.6 | 43.77 | #485 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=power-systems-agent-benchmark-aider · How It Works · Data refreshed daily, snapshot 2026-11-12.