AeroCopilotBench - Tier-2 Ineffective-Action Rate: leaderboard
Metric: Ineffective write attempts (%; lower is better). Source: arxiv.org. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.5 Flash (High) | 5.5 |
| 2 | DeepSeek V4 Pro (High) | 5.7 |
| 3 | Kimi K2.6 (Thinking) | 6.8 |
| 4 | Qwen 3.5 122B A10B (Thinking) | 8.2 |
| 5 | DeepSeek V4 Flash (High) | 10.4 |
| 6 | MiniMax-M2.5 | 11 |
| 7 | Qwen 3.5 397B A17B (Thinking) | 11 |
| 8 | GLM-5.2 (High) | 15 |
| 9 | Nex N2 Pro | 15.7 |
| 10 | GPT-5.6 Sol (High) | 16.8 |
| 11 | GLM-5.1 (Thinking) | 18.9 |
Interactive version: theaggregate.ai/benchmark?slug=aerocopilotbench-tier-2-ineffective-action-rate · How It Works · Data refreshed daily, snapshot 2026-09-19.