← Back

Wire · technology

Artificial Analysis Ships Index v4.3, Adds AutomationBench-AA

Published

8 September 2026

Topic

technology

Sectors

AI Frontier ModelsAI Infrastructure

Source

Read at aiweekly.co

Verified

Fusion42 · 8 September 2026 · Fusion42 review

Artificial Analysis released Intelligence Index v4.3, replacing τ³-Banking with a 657-task AutomationBench-AA benchmark, expanding focus to broader business workflows and upgrading Terminal-Bench to v4.0. GPT-6 Astra leads cost-efficiency, offering 57% lower cost per task compared to Claude Fable 5.1 while tying on top-line score.

This Wire brief sits within Fusion42's coverage of AI Frontier Models and AI Infrastructure.

◆ The Wire takeaway

You face a new standard in evaluating AI that blends breadth and cost efficiency across business tasks. Your AI product development now demands tighter cost management to compete with models like GPT-6 Astra, which cut task costs by over half without losing accuracy.

Coverage

1 source · 8 Sep 2026

Related on Wire

Topics

AI Frontier ModelsAI Infrastructureai-benchmarksautomationcost-efficiencyworkflow-aiindustry-benchmarks