Agentic benchmark
AA AutomationBench leaderboard
Artificial Analysis AutomationBench. Every model the catalog carries a published AA AutomationBench value for, ranked by that value.
CategoryAgentic
MeasureTask success rate
TasksBusiness-process automation tasks
DifficultyAgentic automation
An independently evaluated automation benchmark from Artificial Analysis.
AA AutomationBench ranking
14 models with a published AA AutomationBench value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Task success rate |
|---|---|---|---|
| 1 |  DeepSeek V4.1 Flash | DeepSeek | 68.9 |
| 2 |  GPT-6 Astra | OpenAI | 68.5 |
| 3 |  Grok 4.6 | xAI | 66.7 |
| 4 |  GLM-5.3 | Z.AI | 62.2 |
| 5 | GLM-5.3-Flash | Z.AI | 60.4 |
| 6 | GPT-5.6 Sol | OpenAI | 60.1 |
| 7 |  Gemini 3.8 Flash | 59.9 | |
| 8 | GPT-5.6 Terra | OpenAI | 59.6 |
| 9 |  Claude Fable 5.1 | Anthropic | 59.4 |
| 10 |  Kimi K3 | Moonshot AI | 58.3 |
| 11 |  Muse Spark 1.3 | Meta | 57.9 |
| 12 | DeepSeek V4 Pro 0813 | DeepSeek | 56.7 |
| 13 | Claude Opus 5 | Anthropic | 56.6 |
| 14 | Claude Fable 5 | Anthropic | 54.1 |
Evidence key: Observed
Rows are ordered by the value Artificial Analysis AutomationBench Benchmark Leaderboard published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.