AutomationBench-AA - Operations - Raw Objective Completion: leaderboard
Metric: Raw Objective Completion (%). Source: artificialanalysis.ai. 167 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-6 (Max) | 97.01 |
| 2 | GPT-6 (xHigh) | 96.57 |
| 3 | GPT-5.6 Sol (Max) | 95.38 |
| 4 | GPT-6 (High) | 95.18 |
| 5 | GPT-6 (Medium) | 94.99 |
| 6 | GPT-5.6 Sol (xHigh) | 94.87 |
| 7 | Qwen 3.8 Max (0902) | 94.83 |
| 8 | Qwen3.8-Flash-Next | 94.65 |
| 9 | Claude Opus 5 (Adaptive Reasoning, Max Effort) | 94.37 |
| 10 | Kimi K3 (Max) | 94.24 |
| 11 | Grok 4.6 (High) | 94.24 |
| 12 | Qwen3.8 2.4T A95B | 93.81 |
| 13 | Gemini 3.7 Flash (High) | 93.68 |
| 14 | GLM-5.3 (Max) | 93.64 |
| 15 | Grok 4.6 (Medium) | 93.01 |
Interactive version: theaggregate.ai/benchmark?slug=automationbench-aa-operations-raw-objective-completion · How It Works · Data refreshed daily, snapshot 2026-09-19.