We introduce the AI Productivity Index for Agents (APEX-Agents), a benchmark for assessing whether AI agents can execute long-horizon, cross-application tasks created by investment banking analysts, management consultants, and corporate lawyers. APEX-Agents requires agents to navigate realistic work environments with files and tools. We test eight agents for the leaderboard using Pass@1. Gemini 3 Flash (Thinking=High) achieves the highest score of 24.0%, followed by GPT-5.2 (Thinking=High), Claude Opus 4.5 (Thinking=High), and Gemini 3 Pro (Thinking=High). We open source the APEX-Agents benchmark (n=480) with all prompts, rubrics, gold outputs, files, and metadata. We also open source Archipelago, our infrastructure for agent execution and evaluation.
@article{arxiv.2601.14242,
title = {APEX-Agents},
author = {Bertie Vidgen and Austin Mann and Abby Fennelly and John Wright Stanly and Lucas Rothman and Marco Burstein and Julien Benchek and David Ostrofsky and Anirudh Ravichandran and Debnil Sur and Neel Venugopal and Alannah Hsia and Isaac Robinson and Calix Huang and Olivia Varones and Daniyal Khan and Michael Haines and Austin Bridges and Jesse Boyle and Koby Twist and Zach Richards and Chirag Mahapatra and Brendan Foody and Osvald Nitski},
journal= {arXiv preprint arXiv:2601.14242},
year = {2026}
}