mlsanityData to 14 Oct 2026

APEX-Agents

arXiv 2601.14242 ↗First posted 20 Jan 2026cs.CL cs.AI cs.LGcited by 32

On the map: LLM agents and tool use → General LLM agent frameworks and benchmarks

Abstract (arXiv)

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.

Key facts. Cited by 32 papers in the map (data through 14 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 5 of its 6 references that are in the map. First posted on arXiv 20 Jan 2026.

Builds on

5 of 6 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025ARE: Scaling Up Agent Environments and Evaluations21 Sep
  2. 2025The Measurement Imbalance in Agentic AI Evaluation Undermines Industry Productivity Claims1 Jun

Other references · 3

  1. 2024AI Agents That Mattercited by 219
  2. 2024The BrowserGym Ecosystem for Web Agent Researchcited by 161
  3. 2025The AI Productivity Index (APEX)cited by 10

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 32.

32 papers.

Sort:

By area:

  1. 2026Kimi K3: Open Frontier Intelligencecited by 138
  2. 2026OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Taskscited by 40
  3. 2026Stop Comparing LLM Agents Without Disclosing the Harnesscited by 34
  4. 2026Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Gradingcited by 31
  5. 2026RRSI: Regularized Recursive Self-Improvement of Agent Harnessescited by 13
  6. 2026Evo-Bench: Can Language Models Improve Agent Harness?cited by 11
  7. 2026π-Bench: Evaluating Proactive Personal Assistant Agents in Long-Horizon Workflowscited by 6
  8. 2026One Success Isn't Reliability: Thinkingbox, a Sandbox and Benchmark for Agents in Stateful Business Workflowscited by 3
  9. 2026Locating Hidden Failures Makes Long-Horizon Agents More Reliablecited by 3
  10. 2026OmegaUse-OfficeVal: Benchmarking LLM Agents on Long-Horizon Office-Suite Tasks with Economic Groundingcited by 2
  11. 2026Apodex 1.1: Scaling Agentic Intelligence for Complex Workcited by 2
  12. 2026Cost-Utility Alignment in LLM Agent Trajectories:Profiling,Attribution,Diagnosis,Adaptation,and Evaluationcited by 2
20 more
  1. 2026Is Bash All You Need? An Empirical Study of Tool Interfaces for Enterprise Digital Worker Agentscited by 2
  2. 2026COMPOSITE-Stemcited by 1
  3. 2026Insights Generator: Systematic Corpus-Level Trace Diagnostics for LLM Agentscited by 1
  4. 2026CollabSkill: Evaluating Human-Agent Collaboration On Real-World Taskscited by 1
  5. 2026DuMateBench: Evaluating Autonomous Agents in Complex Real-World Workflowscited by 1
  6. 2026AsymmetryZero: A Framework for Operationalizing Human Expert Preferences as Semantic Evalscited by 0
  7. 2026Relay-Bench: Evaluating LLMs on Multi-Domain Reasoning Chainscited by 0
  8. 2026SQBench: A Benchmark for Evaluating Task Delivery by Language-Model Agents in Production-Oriented Workflowscited by 0
  9. 2026The Reasoning Tax: Token Economics of LLM Reasoning Across Task Types and Deployment Contextscited by 0
  10. 2026Efficient Test-Time Adaptation through Human-AI Interactioncited by 0
  11. 2026xDailyBench: Benchmarking LLMs on Professional Consultation for Real-Life Problemscited by 0
  12. 2026Studying Without a Syllabus: Task-Agnostic Environment Preprocessingcited by 0
  13. 2026Beyond Outcomes: Dual-View Relational Learning for Efficient Agent Benchmarkingcited by 0
  14. 2026After the Fix: Transfer of Corrected Agent Experiencecited by 0
  15. 2026APEX-Voice: Can Voice Agents Complete Professional Workflows Through Full-Duplex Interactioncited by 0
  16. 2026Long-Horizon Scaling: How Model Capabilities Shape the Returns to Computationcited by 0
  17. 2026FinAutoRubric: Expert-Guided Automatic Rubric Generation for Evaluating Financial Research Agentscited by 0
  18. 2026WorkGenesis: Building the Worlds That Teach Agents to Workcited by 0
  19. 2026APEX-Harm: Evaluating Agent Safety in Shared Professional Workspacescited by 0
  20. 2026Prorsa: Unleashing the Potential of Prefill-Decode Disaggregation in Agentic Reinforcement Learningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026BankerToolBench: Evaluating AI Agents in End-to-End Investment Banking Workflows
  2. 2026Do Agent Benchmarks Measure Capability? Protocol Validity in the Age of Agentic AI
  3. 2025Holistic Agent Leaderboard: The Missing Infrastructure for AI Agent Evaluation
  4. 2025Towards Outcome-Oriented, Task-Agnostic Evaluation of AI Agents
  5. 2026EnterpriseClawBench: Benchmarking Agents from Real Workplace Sessions
  6. 2026AlphaEval: Evaluating Agents in Production
  7. 2025Establishing Best Practices for Building Rigorous Agentic Benchmarks
  8. 2026Evaluating Deep Research Agents on Expert Consulting Work: A Benchmark with Verifiers, Rubrics, and Cognitive Traps