AutomationBench-AA - Support - Raw Objective Completion: leaderboard
Artificial Analysis's implementation of Zapier's AutomationBench restricted to its Support tasks: the share of task objectives an agent completes in simulated SaaS applications, counted before the guardrail-violation adjustment used by the AutomationBench-AA headline. Reported 0-100 (%); higher is better.
Metric: Raw Objective Completion (%). Source: artificialanalysis.ai. Saturation forecast: Around February 2027. 243 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Argon (High) | 94.72 | #9 (Gemini 4 Argon) |
| 2 | Gemini 4 Carbon (High) | 94.26 | #1 (Gemini 4 Carbon) |
| 3 | GPT-6 Astra (xHigh) | 92.43 | #6 (GPT-6 Astra) |
| 4 | Muse Wonder Full (Max) | 92.3 | #12 (Muse Wonder Full) |
| 5 | GPT-6.1 Sol (High) | 92 | #7 (GPT-6.1 Sol) |
| 6 | GPT-6 Astra (Max) | 91.93 | #6 (GPT-6 Astra) |
| 7 | GPT-6.1 Sol (Max) | 91.84 | #7 (GPT-6.1 Sol) |
| 8 | Kimi K3.1 Pro | 91.59 | #17 |
| 9 | GPT-6 Astra (High) | 91.34 | #6 (GPT-6 Astra) |
| 10 | DeepSeek V4.1 Flash (1017) (High) | 91.31 | #27 (DeepSeek V4.1 Flash (1017)) |
| 11 | Grok 4.7 (High) | 90.97 | #26 (Grok 4.7) |
| 12 | MiniMax-M3.1-Flash | 90.08 | #52 |
| 13 | GPT-6.1 Sol (xHigh) | 90.03 | #7 (GPT-6.1 Sol) |
| 14 | GPT-6 Sol (Max) | 89.76 | #19 (GPT-6 Sol) |
| 15 | Muse Wonder (Max) | 89.76 | #18 (Muse Wonder) |
No result here: #3 Claude Opus 5.5, #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=automationbench-aa-support-raw-objective-completion · How It Works · Data refreshed daily, snapshot 2026-11-05.