DeliverBench: Benchmarking AI Agents on Delivering Realistic Multi-Stage Professional Projects
cs.AI
Submitted: 2026-11-02
Updated: 2026-11-09
Terminology
Sources
- DAYJOB: A Benchmark for Long-Horizon Professional Work
- AI Agents That Matter
- JobBench: Aligning Agent Work With Human Will
- AutomationBench
- PaperBench: Evaluating AI's Ability to Replicate AI Research
- SQBench: A Benchmark for Evaluating Task Delivery by Language-Model Agents in Production-Oriented Workflows
- Agents' Last Exam
- Workspace-Bench 1.0: Benchmarking AI Agents on Workspace Tasks with Large-Scale File Dependencies
- APEX-Agents
- OfficeBench: Benchmarking Language Agents across Multiple Applications for Office Automation
- $\tau$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains
- OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks
- $\pi$-Bench: Evaluating Proactive Personal Assistant Agents in Long-Horizon Workflows
- Agent-as-a-Judge: Evaluate Agents with Agents
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents