Woofun AI reports that Artificial Analysis has introduced six industry-specific indices to evaluate real-world business performance across finance, law, healthcare, operations, engineering, and economics. These indices utilize ONET Job Activity Classification weights to simulate actual task distributions, combining large model capability tests with industry knowledge bases. Initial results show Claude Fable 5 ranking first in all eight indices, while open-source GLM-5.2 achieved five top positions among the industry metrics, scoring 53 in engineering, close to Claude Sonnet 5's 55. The assessment highlighted significant efficiency disparities: Claude Fable 5 incurred a $3.48 single-task cost in operations, over 100 times higher than DeepSeek V4 Pro's $0.03, whereas Gemini 3.1 Pro preview operated nearly seven times faster than Claude Fable 5 in legal tasks with only an 11-point score difference.