ITBench-AA: Frontier Models Score Below 50% on the First Benchmark for Agentic Enterprise IT Tasks — by Artificial Analysis and IBM

The ITBench-AA benchmark evaluates frontier AI models on enterprise IT tasks. Current models score below 50%, indicating significant room for improvement. This benchmark highlights challenges in applying AI to complex IT workflows.

WeekAI AgentsHigh-signal source

Signal trust

High-signal sourceSingle sourceEarly signal

Original article excerpt

Server-side extracted preview paragraphs from the original source.

A Blog post by IBM Research on Hugging Face

The source page did not expose a stable readable article body, so only the AI summary is available for now.

Opening the briefing

ITBench-AA: Frontier Models Score Below 50% on the First Benchmark for Agentic Enterprise IT Tasks — by Artificial Analysis and IBM

Original article excerpt

Related stories

Opening the briefing

ITBench-AA: Frontier Models Score Below 50% on the First Benchmark for Agentic Enterprise IT Tasks — by Artificial Analysis and IBM

Original article excerpt

Related stories

Anthropic releases Opus 4.8 with new ‘dynamic workflow’ tool

How Endava builds an agentic organization with Codex

Replit’s vibe coding platform just got a Visa-backed identity layer for AI agents — and it changes how agents spend money

Sesame, the conversational AI startup from Oculus founders, launches its iOS app