EnergyAgentBench is a new benchmark with 70 task variants that evaluates LLM agents on live energy data for datacenter siting, long-horizon optimization, and causal grid diagnosis.
super hub Canonical reference
AgentBench: Evaluating LLMs as Agents
Canonical reference. 86% of citing Pith papers cite this work as background.
172
Pith papers citing it
Background
86% of classified citations
abstract
The potential of Large Language Model (LLM) as agents has been widely acknowledged recently. Thus, there is an urgent need to quantitatively \textit{evaluate LLMs as agents} on challenging tasks in interactive environments. We present AgentBench, a multi-dimensional benchmark that consists of 8 distinct environments to assess LLM-as-Agent's reasoning and decision-making abilities. Our extensive test over \num API-based and open-sourced (OSS) LLMs shows that, while top commercial LLMs present a strong ability of acting as agents in complex environments, there is a significant disparity in performance between them and many OSS competitors that are no larger than 70B. We identify the typical reasons of failures in environments and LLMs, showing that poor long-term reasoning, decision-making, and instruction following abilities are the main obstacles for developing usable LLM agents. Improving instruction following and training on high quality multi-round alignment data could improve agent performance. And different from existing assumptions, training on code present ambivalent impacts on different agent tasks. Datasets, environments, and an integrated evaluation package for AgentBench are released at https://github.com/THUDM/AgentBench.
hub tools
citation-role summary
background 38
dataset 5
baseline 1
citation-polarity summary
claims ledger
- abstract The potential of Large Language Model (LLM) as agents has been widely acknowledged recently. Thus, there is an urgent need to quantitatively \textit{evaluate LLMs as agents} on challenging tasks in interactive environments. We present AgentBench, a multi-dimensional benchmark that consists of 8 distinct environments to assess LLM-as-Agent's reasoning and decision-making abilities. Our extensive test over \num API-based and open-sourced (OSS) LLMs shows that, while top commercial LLMs present a strong ability of acting as agents in complex environments, there is a significant disparity in perfo
authors
co-cited works
representative citing papers
A new native-runtime benchmark reveals that current frontier AI agents succeed on at most 62 percent of realistic long-horizon CLI tasks.
SimWorld Studio deploys an evolving coding agent to create adaptive 3D environments that co-evolve with embodied learners, delivering 18-point success-rate gains over fixed environments in navigation benchmarks.
PhysicianBench is a new benchmark of 100 physician-reviewed, execution-grounded tasks in live EHR environments where the best LLM agent reaches only 46% success and open-source models reach 19%.
AgentDojo introduces an extensible evaluation framework populated with realistic agent tasks and security test cases to measure prompt injection robustness in tool-using LLM agents.
OSWorld provides the first unified real-computer benchmark for open-ended multimodal agent tasks, exposing large performance gaps between humans and state-of-the-art LLM/VLM agents.
LongBench is the first bilingual multi-task benchmark for long context understanding in LLMs, containing 21 datasets in 6 categories with average lengths of 6711 words (English) and 13386 characters (Chinese).
MultiUAV-Plat supplies a new RESTful simulation platform and 1500-task benchmark where Agent4Drone reaches 57.9% task pass rate versus 30.6% for ReAct baseline across 75 multi-UAV missions.
PrincipalBench exposes a sharp split in frontier LLMs between selective and over-refusing behavior on multi-party loyalty, with prompt scaffolding and KL distillation reducing harm rates but only along an existing leak/over-refusal trade-off.
SpreadsheetBench 2 provides 321 expert-validated tasks from authentic business data showing frontier LLMs reach only 34.89% overall accuracy on end-to-end spreadsheet workflows.
CLQT is a new closed-loop, cost-aware benchmark that diagnoses LLM trading agent capabilities through strategy-consistent metrics and hash-verifiable trails rather than outcome rankings.
LLM agents often fail to abstain at the right time in uncertain multi-turn tasks, and the CONVOLVE context engineering method raises timely abstention rates on WebShop from 26.7 to 57.4 without parameter updates.
RigorBench evaluates AI coding agents on process discipline via five pillars and reports 41% higher process scores and 17% better outcome correctness with structured approaches on 30 tasks.
OR-Space is a benchmark for LLM agents performing full-lifecycle optimization tasks across Build, Revise, and Explain modes in executable multi-artifact workspaces.
JobBench is a new benchmark with 130 occupational tasks where the best of 36 tested AI models achieves only 45.9% success.
ScaleWoB generates 100+ synthetic interactive GUI environments and 1000+ verifiable tasks as web pages, releasing a 120-task mobile benchmark where state-of-the-art agents achieve 27.92% success (17.82% on long-horizon tasks) versus 92.08% for humans, with synthetic results generalizing to real apps
Introduces QGP and PushBench to evaluate LLM agent persistence on quantitative goals, showing specialized controllers outperform baselines on verifier-checked artifact collection tasks.
Introduces MCP-TDP benchmark showing near-100% attack success on models like GPT-4o for tool description poisoning and proposes reactive self-correction defense.
DART is a modular runtime that certifies semantically recoverable boundaries for failed tool-agent instances and selects admissible restore points that preserve downstream commitments or blocks recovery.
Boiling the Frog is a new stateful multi-turn benchmark that finds an aggregate 44.4% strict attack success rate for incremental safety violations across nine AI models, with rates ranging from 20.5% to 92.9%.
PrefBench benchmark shows zero-shot LLMs achieve deal rates above 0.99 but seller profits only slightly above random and far below a simple concession heuristic across 7,500 episodes.
SCICONVBENCH is a new benchmark evaluating LLMs on multi-turn disambiguation and inconsistency resolution for task formulation in computational science, with frontier models reaching only 52.7% success on fluid mechanics disambiguation cases.
BioXArena benchmarks LLM agents on generating end-to-end ML pipelines for 76 multi-modal biomedical tasks, with MLEvolve plus Gemini-3.1-Pro scoring highest at 0.666.
π-Bench is a new benchmark for evaluating proactive personal assistant agents on 100 multi-turn tasks that include hidden intents, inter-task dependencies, and cross-session continuity.
citing papers explorer
-
AIT Academy: Cultivating the Complete Agent with a Confucian Three-Domain Curriculum
AIT Academy introduces a tripartite curriculum for AI agents across natural science, humanities, and social science domains, with reported gains of 15.9 points in security and 7 points in social reasoning under specific scheduling.
-
SLALOM: Simulation Lifecycle Analysis via Longitudinal Observation Metrics for Social Simulation
SLALOM uses phase constraints called gates and dynamic time warping to quantitatively measure whether simulated social trajectories follow empirically plausible paths instead of just reaching correct end states.
-
ForkKV: Scaling Multi-LoRA Agent Serving via Copy-on-Write Disaggregated KV Cache
ForkKV uses copy-on-write disaggregated KV cache with DualRadixTree and ResidualAttention kernels to deliver up to 3x throughput over prior multi-LoRA serving systems with negligible quality loss.
-
Do Agent Societies Develop Intellectual Elites? The Hidden Power Laws of Collective Cognition in LLM Multi-Agent Systems
LLM agent societies develop power-law coordination cascades and intellectual elites through an integration bottleneck that grows with system size.
-
Beyond Benchmark Islands: Toward Representative Trustworthiness Evaluation for Agentic AI
Defines agentic trustworthiness via five properties and proposes HAAF, a scenario-distribution framework with a Trustworthy Optimization Factory that transfers interventions across 13 models from seven families on a 100-scenario suite.
-
Characterizing Faults in Agentic AI: A Taxonomy of Types, Symptoms, and Root Causes
An empirical study of real-world issues yields a taxonomy of 34 fault types, symptoms, and root causes in agentic AI systems, validated by 145 practitioners.
-
SoK: Agentic Skills -- Beyond Tool Use in LLM Agents
The paper systematizes agentic skills beyond tool use, providing design pattern and representation-scope taxonomies plus security analysis of malicious skill infiltration in agent marketplaces.
-
NeuroSymActive: Differentiable Neural-Symbolic Reasoning with Active Exploration for Knowledge Graph Question Answering
NeuroSymActive combines soft-unification symbolic modules, a neural path evaluator, and Monte-Carlo-style active exploration to reach strong answer accuracy on KGQA benchmarks while cutting graph lookups and model calls versus standard retrieval baselines.
-
Learning to Configure Agentic AI Systems
ARC learns per-query agent configurations via a lightweight hierarchical SMDP policy, delivering 31.3% higher reasoning accuracy, 13.95% higher tool-use accuracy, and doubled success on an agent benchmark compared to budget-matched baselines.
-
AgentXRay: White-Boxing Agentic Systems via Workflow Reconstruction
AgentXRay formulates workflow reconstruction as combinatorial optimization and uses Monte Carlo Tree Search with Red-Black Pruning to approximate black-box agent behaviors via output-based proxy metrics.
-
MirrorBench: A Benchmark to Evaluate Conversational User-Proxy Agents for Human-Likeness
MirrorBench defines a reproducible benchmark combining lexical metrics (MATTR, Yule's K, HD-D) and LLM-judge metrics with calibration controls to measure human-likeness of user-proxy agents across four datasets.
-
CryptoBench: A Dynamic Benchmark for Expert-Level Evaluation of LLM Agents in Cryptocurrency
CryptoBench is a new dynamic benchmark for LLM agents in cryptocurrency that reveals a retrieval-prediction imbalance in model performance.
-
Exploring the Secondary Risks of Large Language Models
Introduces secondary risks as a new class of LLM failures from benign prompts, defines two primitives, proposes SecLens search framework, and releases SecRiskBench showing risks are widespread across 16 models.
-
DeepResearch Bench: A Comprehensive Benchmark for Deep Research Agents
DeepResearch Bench supplies 100 expert-crafted PhD-level tasks and two human-aligned evaluation frameworks to measure deep research agents on report quality and citation accuracy.
-
Training Language Models to Self-Correct via Reinforcement Learning
SCoRe uses multi-turn online RL with regularization on self-generated traces to improve LLM self-correction, achieving 15.6% and 9.1% gains on MATH and HumanEval for Gemini models.
-
WebCanvas: Benchmarking Web Agents in Online Environments
WebCanvas creates a dynamic benchmark for web agents with a noise-resistant evaluation metric, the Mind2Web-Live dataset of 542 tasks, and open-source tools and agent framework for ongoing online testing.
-
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
LiveCodeBench collects 400 recent contest problems to create a contamination-free benchmark evaluating LLMs on code generation and related capabilities like self-repair and execution.
-
A Survey on Large Language Model based Autonomous Agents
A survey of LLM-based autonomous agents that proposes a unified framework for their construction and reviews applications in social science, natural science, and engineering along with evaluation methods and future directions.
-
MAS-Lab: A Specification-Driven Validation Framework for Reliable Multi-Agent Systems
MAS-Lab proposes a specification-driven framework with Spec, MAS-OS, and Labs layers to enable intent-based validation and reliable evolution of multi-agent systems.
-
RESOURCE2SKILL: Distilling Executable Agent Skills from Human-Created Multimodal Resources
RESOURCE2SKILL converts multimodal human resources into a hierarchical Skill Wiki of executable agent skills, reporting +11.9 percentage point average gains over no-skill baselines across seven authoring domains.
-
Direct Causation in International Humanitarian Law and the Challenge of AI-Mediated Civilian Cyber Operations
Autonomous AI cyber systems deployed by civilians fail the one-causal-step and integral-part requirements of the IHL direct participation test because harm arises from post-disengagement system decisions.
-
DeXposure-Claw: An Agentic System for DeFi Risk Supervision
DeXposure-Claw combines a graph time-series foundation model for forecasting DeFi networks with rule-based monitors and data-health gates to emit regulator-aligned risk tickets, evaluated via a new six-axis benchmark on five years of real weekly data.
-
Testing Agentic Workflows with Structural Coverage Criteria
Introduces coordination-graph structural coverage for multi-agent workflows with DSPy-generated tests, evaluated on ten SDK benchmarks covering 49 agents and 403 obligations.
-
CausalFlow: Causal Attribution and Counterfactual Repair for LLM Agent Failures
CausalFlow introduces an interventional causal framework that attributes failures in LLM agent traces via counterfactual interventions and generates minimal repairs, validated on four benchmarks.
-
Code as Agent Harness
A survey that organizes existing work on LLM-based agents around code as the central harness, structured in three layers of interfaces, mechanisms, and multi-agent scaling, with applications across domains and listed open challenges.
-
Latent Action Reparameterization for Efficient Agent Inference
LAR learns a compact latent action space from trajectories that shortens the effective decision horizon for LLM agents, reducing token count and inference time while preserving task success.
-
Beyond Scaling: Agents Are Heading to the Edge
Personal agents require edge deployment to preserve high-fidelity local context and zero-latency loops, as claimed through three structural shifts away from cloud-centric designs.
-
The Alpha Illusion: Reported Alpha from LLM Trading Agents Should Not Be Treated as Deployment Evidence
Reported alpha from end-to-end LLM trading agents does not constitute deployment evidence until it passes structural tests for temporal integrity, frictions, robustness, calibration, execution, and disaggregation.
-
Runtime-Structured Task Decomposition for Agentic Coding Systems
Runtime-structured task decomposition reduces retry costs in agentic coding systems by up to 51.7% versus monolithic prompts by rerunning only failed subtasks on two software engineering workloads.
-
The Misattribution Gap: When Memory Poisoning Looks Like Model Failure in Agentic AI Systems
Memory poisoning via lost-provenance documents in agent memory stores creates agent misconduct that safety systems misattribute to model failure; the paper defines Semantic Norm Drift, releases a benchmark, and proposes a new testing method plus a defense.
-
Designing Intelligent Enterprise Agents: A Capability-Aligned Multi-Agent Architecture
CEAD architecture for intelligent enterprise agents achieves 70.6% safe success rate on 10,000 tasks by making agent design the primary abstraction rather than governance.
-
From Agent Loops to Deterministic Graphs: Execution Lineage for Reproducible AI-Native Work
Execution lineage models AI-native work as a DAG of computations with explicit dependencies, achieving perfect state preservation in controlled update tasks where loop-based agents introduce churn and contamination.
-
Complexity Horizons of Compressed Models in Analog Circuit Analysis
Prerequisite graphs map compressed LLM performance boundaries in analog circuit analysis to allow selecting the smallest viable model for a given task complexity.
-
Toward a Science of Intent: Closure Gaps and Delegation Envelopes for Open-World AI Agents
Intent compilation turns vague human goals into verifiable artifacts, using closure-gap vectors and delegation envelopes to separate open-world agent challenges from closed-world solvers and to benchmark closure fixes against extra search.
-
Layered Mutability: Continuity and Governance in Persistent Self-Modifying Agents
Persistent self-modifying AI agents exhibit compositional drift from mismatches across five mutability layers, with governance difficulty rising under rapid mutation, strong coupling, weak reversibility, and low observability, as indicated by a 0.68 identity hysteresis ratio in a preliminary ratchet
-
Spec Kit Agents: Context-Grounded Agentic Workflows
A multi-agent SDD framework with phase-level context-grounding hooks improves LLM-judged quality by 0.15 points and SWE-bench Lite Pass@1 by 1.7 percent while preserving near-perfect test compatibility.
-
An Empirical Study of Multi-Agent Collaboration for Automated Research
Subagent architectures deliver stable high-throughput optimization under tight time limits while agent teams enable deeper refactoring at the cost of higher fragility.
-
From Pixels to Digital Agents: An Empirical Study on the Taxonomy and Technological Trends of Reinforcement Learning Environments
An empirical literature analysis reveals a bifurcation in RL environments into Semantic Prior (LLM-dominated) and Domain-Specific Generalization ecosystems with distinct cognitive fingerprints.
-
AgentCrypt: Advancing Privacy and (Secure) Computation in AI Agent Collaboration
AgentCrypt introduces a deterministic three-tier privacy framework for AI agent collaboration that uses masking and homomorphic encryption to protect data independently of model accuracy.
-
Towards Understanding, Analyzing, and Optimizing Agentic AI Execution: A CPU-Centric Perspective
The paper analyzes CPU bottlenecks in agentic AI serving, selects representative workloads, and demonstrates that CPU-aware scheduling optimizations COMB and MAS can reduce P50 latency by up to 1.7x and total latency by up to 2.49x on two hardware systems.
-
A Comprehensive Survey of Agents for Computer Use: Foundations, Challenges, and Future Directions
A survey of 87 agents for computer use and 33 datasets that introduces a three-dimensional taxonomy across domain, interaction, and agent perspectives and identifies six research gaps.
-
AppAgent: Multimodal Agents as Smartphone Users
AppAgent lets large language models operate diverse smartphone apps via visual interactions and learns app usage from exploration or demonstrations.
-
What makes a harness a harness: necessary and sufficient conditions for an agent harness
Proposes and tests a constitutive definition of 'agent harness' via conceptual analysis of literature and six real systems.
-
Skill Availability and Presentation Granularity in Large-Language-Model Agents: A Controlled SkillsBench Study
In a 30-task SkillsBench study, skill availability boosts GPT-5.5 and DeepSeek V4-Flash agent pass rates substantially while presentation-granularity variations yield small uncertain effects.
-
Grimlock: Guarding High-Agency Systems with eBPF and Attested Channels
Grimlock is a guard system that moves trust enforcement into the sandbox substrate for high-agency agents via eBPF interception and attested TLS channels.
-
Evaluating Large Language Models as Live Strategic Agents: Provider Performance, Hybrid Decomposition, and Operational Gaps in Timed Risk Play
Gemini-3.1-pro-preview won 20 of 32 Risk games through superior objective tracking and execution conversion, while a hybrid test with fixed execution showed near-equal planner performance across providers.
-
Rethinking Code Review in the Age of AI: A Vision for Agentic Code Review
Proposes a five-stage agentic AI framework for code review with human quality gates to maintain context, accountability, and team understanding.
-
ChromaFlow: A Negative Ablation Study of Orchestration Overhead in Tool-Augmented Agent Evaluation
ChromaFlow reports a negative ablation in which expanded orchestration on GAIA Level-1 tasks reduced accuracy and increased tracebacks, timeouts, and token costs.
-
Social Theory Should Be a Structural Prior for Agentic AI: A Formal Framework for Multi-Agent Social Systems
Agentic AI needs social theory as structural priors in the MASS framework to model emergent dynamics from multi-agent interactions.
-
Stop Comparing LLM Agents Without Disclosing the Harness
The Binding Constraint Thesis states that harness configuration governs performance variance more than model choice in long-horizon agent tasks, leading to misattribution in evaluations.