[Grok 4.5 Tops SaaS Workflow Evaluation, Costs About One-Fourth of Claude] According to monitoring by Beating, independent evaluation agency Artificial Analysis has updated the AutomationBench-AA rankings. SpaceXAI's Grok 4.5 scored 51.4%, taking the top spot, surpassing Claude Fable 5's 48.6% and Claude Opus 4.8's 48.5%. AutomationBench-AA is based on Zapier's AutomationBench and tests whether AI agents can complete real SaaS workflows according to enterprise rules. The test set includes 657 multi-step tasks, covering 40 simulated application environments, including Gmail, Google Sheets, Slack, Salesforce, and HubSpot. The scoring criteria are stringent. Models not only need to complete tasks but must also avoid crossing enterprise-defined red lines. If a task triggers a violation, it is scored as 0. Grok 4.5's raw task completion rate is close to 80%, higher than Claude Fable 5 and Claude Opus 4.8. Even after red-line deductions, it still ranks first with a score of 51.4%, making it the first model to exceed 50% in this evaluation. Its cost advantage is even more pronounced. Grok 4.5's cost per task is $0.34, lower than Claude Fable 5's $1.35, Claude Opus 4.8's $1.46, and GPT-5.5 xhigh's $1.28. However, Grok 4.5 is not the most stable model. On average, it triggers 0.63 violations per task, higher than Claude Opus 4.8's 0.55 and Gemini 3.5 Flash's 0.46. [Original Link]