Loading leaderboard data…
AutomationBench-AA: Agentic SaaS Workflow Benchmark
As of 29 Oct 2026 · 236 ranked rows · Original source ↗
Cost per Intelligence Index task uses Artificial Analysis profile values captured 29 Oct 2026, weighted across the Index evaluations. Workbook cost values mentioned in the source notes below are evaluation-specific.
AA Cost Chart Label is the displayed source value; Cost per Task (USD) is numeric. A bound such as <0.1¢ remains in the label while its numeric cell is blank; if both cells are blank, AA showed no plotted cost point for this model. GPT-6.1 Sol effort-variant rows use scores from Artificial Analysis model profiles captured 2026-09-30; insertion ranks are documented in docs/GPT-6.1-SOL-RANK-IMPACT.md. Values refreshed from the Artificial Analysis evaluation page captured 2026-10-29.
5 results
| 5 | Anthropic | Claude Sonnet 5.5 (Adaptive Reasoning, Max Effort, Default Fallback) | 0.717688 | Sep 2026 | $10.02 |
| 7 | Anthropic | Claude Opus 5.5 (Adaptive Reasoning, Max Effort, Default Fallback) | 0.695387 | Sep 2026 | $10.04 |
| 45 | Anthropic | Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback) | 0.593759 | Sep 2026 | $13.04 |
| 70 | Anthropic | Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) | 0.540713 | Jun 2026 | — |
| 118 | Anthropic | Claude Haiku 5.5 (Adaptive Reasoning, Max Effort, Default Fallback) | 0.35411 | Oct 2026 | $0.41 |