APEX-Agents
A benchmark that evaluates whether AI agents can complete long-horizon, cross-application professional tasks in investment banking, consulting, and corporate law.
APEX-Agents evaluates agentic AI systems on realistic, multi-step professional workflows rather than single-turn prompts. It comprises 33 data-rich simulated work environments with 480 tasks requiring agents to navigate complex file systems, work across documents, spreadsheets, PDFs, chat, email, and calendar. The benchmark measures Pass@1 — probability of success on the first attempt. The entire dataset is open-sourced on Hugging Face, along with the Archipelago Docker-based evaluation harness on GitHub.
Use cases
Evaluating enterprise AI agent readiness
Benchmarking agentic systems on long-horizon workflows
Testing agent reliability under realistic conditions
Guiding RL training strategies
Integrations
Hugging Face
GitHub (Archipelago)
Docker
Expected outcomes
Standardized agent evaluation before deployment
Identification of failure modes
Improved agent training
Reviews
GCC buyer feedback
Reviews are written by GCC buyers and published after moderation.
No reviews yet
Buyer reviews will appear here once published.
Company
Mercor
San Francisco, USA
Founded 2023
Primary Verticals
Financial Services (BFSI)Healthcare & Life SciencesManufacturingRetail & E-commerceTelecom & Media
Functions
Data & AnalyticsLegal & ComplianceFinance & Accounting
Industries
Financial Services (BFSI)
3
Integrations
4
Use cases
Is this your company? Claim & customize your profile
This profile was created using publicly available information.