Institutions / HKU

HKU

The University of Hong Kong

www.hku.hk · @HKUniversity

Their benchmarks in the atlas

• BIRD-Interactby HKU and Google CloudTask level · Success rate, a-Interact, Full, both phases

  22.7%MERIT + GPT-5.4 (GPT-4o user simulator) · 22 May 2026

  Emerging
• OSWorld 2.0by HKU and Snorkel AIProject level · Binary completion rate (full set, 500 steps)

  44.3%Claude Opus 5 (max, batch tool) · 17 Sep 2026

  Emerging
• OSWorld-Verifiedby HKU, Salesforce AI Research, Carnegie Mellon University and University of WaterlooTask level · Success rate

  90.2%Intelligence-Indeed Agent · 25 Jul 2026

  Saturated
• Spider 2.0by HKU and Salesforce AI ResearchTask level · Execution accuracy (Spider 2.0-Snow)

  96.7%Genloop's Sentinel Agent v2 Pro · 1 Mar 2026

  Saturated

Work their benchmarks cover

Data and analyticsSoftware engineeringOffice and administrative supportDesign, media, and writingScience and researchArchitecture and engineeringFinance and accountingHealthcareDevOps, SRE, and IT operations

<a href="https://canagentswork.com/institutions/hku/"><img src="https://canagentswork.com/og/institutions-hku.png" width="600" height="315" alt="HKU's agent benchmarks, mapped onto the world of work:" loading="lazy"></a>

[![HKU's agent benchmarks, mapped onto the world of work:](https://canagentswork.com/og/institutions-hku.png)](https://canagentswork.com/institutions/hku/)

<a href="https://canagentswork.com/institutions/hku/">Featured in the Work Atlas — Can Agents Work?</a>