# ReliabilityBench: Evaluating LLM Agent Reliability Under Production-Like Stress Conditions

arXiv: https://arxiv.org/abs/2601.06112 · HTML: https://mlsanity.com/paper/2601.06112/

## Key facts

- Cited by 36 papers in the map (data through 14 Oct 2026).
- Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use.
- Builds on 16 of its 36 references that are in the map.
- First posted on arXiv 3 Jan 2026.

## Placed in

- LLM agents and tool use / [Software engineering coding agents](https://mlsanity.com/agents/software-engineering-coding-agents/)

## Builds on (most cited first, 16 of 16)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [Toolformer: Language Models Can Teach Themselves to Use Tools](https://mlsanity.com/paper/2302.04761/) (2023)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs](https://mlsanity.com/paper/2307.16789/) (2023)
- [τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains](https://mlsanity.com/paper/2406.12045/) (2024)
- [AgentBench: Evaluating LLMs as Agents](https://mlsanity.com/paper/2308.03688/) (2023)
- [Beyond Accuracy: Behavioral Testing of NLP models with CheckList](https://mlsanity.com/paper/2005.04118/) (2020)
- [Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and Entailment](https://mlsanity.com/paper/1907.11932/) (2019)
- [BERT-ATTACK: Adversarial Attack Against BERT Using BERT](https://mlsanity.com/paper/2004.09984/) (2020)
- [ToolAlpaca: Generalized Tool Learning for Language Models with 3000 Simulated Cases](https://mlsanity.com/paper/2306.05301/) (2023)
- [DeepXplore: Automated Whitebox Testing of Deep Learning Systems](https://mlsanity.com/paper/1705.06640/) (2017)
- [StableToolBench: Towards Stable Large-Scale Benchmarking on Tool Learning of Large Language Models](https://mlsanity.com/paper/2403.07714/) (2024)
- [Taxonomy of Real Faults in Deep Learning Systems](https://mlsanity.com/paper/1910.11015/) (2019)
- [Metamorphic Testing: A New Approach for Generating Next Test Cases](https://mlsanity.com/paper/2002.12543/) (2020)
- [Chaos Engineering](https://mlsanity.com/paper/1702.05843/) (2017)

## Built on by (most cited first, 30 of 36)

- [When Tools Fail: Benchmarking Dynamic Replanning and Anomaly Recovery in LLM Agents](https://mlsanity.com/paper/2606.05806/) (2026)
- [From Question Answering to Task Completion: A Survey on Agent System and Harness Design](https://mlsanity.com/paper/2606.20683/) (2026)
- [Self-Healing Agentic Orchestrators for Reliable Tool-Augmented Large Language Model Systems](https://mlsanity.com/paper/2606.01416/) (2026)
- [Compiling Agentic Workflows into LLM Weights: Near-Frontier Quality at Two Orders of Magnitude Less Cost](https://mlsanity.com/paper/2605.22502/) (2026)
- [When Agents Disagree With Themselves: Behavioral Consistency as an Uncertainty Signal for LLM Agents](https://mlsanity.com/paper/2602.11619/) (2026)
- [The Devil Is in the Interface: Evaluating How Tool Architecture Shapes Coding Agent Behavior](https://mlsanity.com/paper/2608.11386/) (2026)
- [Where Does Exactly-Once Live? Model, Harness, and Tool-Contract Effects on Duplicate Side Effects in LLM Agents](https://mlsanity.com/paper/2609.29095/) (2026)
- [In-Context Prompting Obsoletes Agent Orchestration for Procedural Tasks](https://mlsanity.com/paper/2604.27891/) (2026)
- [A Self-Healing Framework for Reliable LLM-Based Autonomous Agents](https://mlsanity.com/paper/2605.06737/) (2026)
- [How Much Coordination Gain Is Real? A Paired Noise-Floor Protocol for Multi-Agent LLM Benchmarks](https://mlsanity.com/paper/2606.20695/) (2026)
- [Tangent: An Empirical Study of Testing Practices for LLM-Based Agent Applications](https://mlsanity.com/paper/2608.08413/) (2026)
- [Retry, Switch, or Abstain? Learning Strategy-Aware Tool-Use Policies via Controlled Error Injection](https://mlsanity.com/paper/2608.11977/) (2026)
- [Cost-Utility Alignment in LLM Agent Trajectories:Profiling,Attribution,Diagnosis,Adaptation,and Evaluation](https://mlsanity.com/paper/2608.26195/) (2026)
- [Latency-Quality Routing for Functionally Equivalent Tools in LLM Agents](https://mlsanity.com/paper/2605.14241/) (2026)
- [PlanningBench: Generating Scalable and Verifiable Planning Data for Evaluating and Training Large Language Models](https://mlsanity.com/paper/2605.20873/) (2026)
- [Toward User-Conditioned Evaluation of Personal LLM Agents under Temporal Interventions](https://mlsanity.com/paper/2607.21635/) (2026)
- [Towards an Agent Operating System - Lessons from Classical and Cloud OS](https://mlsanity.com/paper/2607.25076/) (2026)
- [SkillSentry: Reliable Skill Execution for LLM Agents via Runtime Assurance](https://mlsanity.com/paper/2608.09253/) (2026)
- [When Agentic Executions Fail: Detecting and Localizing Runtime Faults from Telemetry](https://mlsanity.com/paper/2608.14680/) (2026)
- [SilentProbe: Measuring Silent Failure in Production APIs Used as Agent Tools](https://mlsanity.com/paper/2609.00035/) (2026)
- [Failure-Transparent Agents: Benchmarking Post-Failure Reporting in Tool-Using Language Models](https://mlsanity.com/paper/2609.35732/) (2026)
- [UndoBench: Separating Task Competence from Recovery Capability in Tool-Using AI Agents](https://mlsanity.com/paper/2610.05622/) (2026)
- [Toward Self-Evolution-Ready Workflow Harnesses: A Reversible Migration Path and Convertibility Taxonomy for Expert LLM Pipelines](https://mlsanity.com/paper/2606.24598/) (2026)
- [When the Database Fails: Prompting LLM Dialogue Agents for Safe Recovery in Task-Oriented Dialogue](https://mlsanity.com/paper/2606.31307/) (2026)
- [Beyond Static Evaluation: Building Simulation Environments for Scalable Agentic Reinforcement Learning](https://mlsanity.com/paper/2607.05773/) (2026)
- [LayerRAG-Bench: A Cross-Layer Reliability Benchmark for Agentic Retrieval-Augmented Generation](https://mlsanity.com/paper/2607.27353/) (2026)
- [No Task Fails Every Time: Why One-Shot Audits Are Structurally Blind to Agent Damage](https://mlsanity.com/paper/2608.15286/) (2026)
- [Towards Risk-free AI Agent Deployment](https://mlsanity.com/paper/2608.16411/) (2026)
- [Same Patient, Different Order: Action-Level Reliability of Clinical LLM Agents Under Repeated Runs](https://mlsanity.com/paper/2609.13582/) (2026)
- [When Hiring Becomes Agent-Mediated: Evaluating Access and Recurrence in Two-Agent R'esum'e Screening](https://mlsanity.com/paper/2609.19530/) (2026)

## Abstract (arXiv)

Existing benchmarks for tool-using LLM agents primarily report single-run success rates and miss reliability properties required in production. We introduce ReliabilityBench, a benchmark for evaluating agent reliability across three dimensions: (i) consistency under repeated execution using passk, (ii) robustness to semantically equivalent task perturbations at intensity ε, and (iii) fault tolerance under controlled tool/API failures at intensity λ. ReliabilityBench contributes a unified reliability surface R(k,ε,λ), textitaction metamorphic relations that define correctness via end-state equivalence rather than text similarity, and a chaos-engineering-style fault injection framework (timeouts, rate limits, partial responses, schema drift). We evaluate two models (Gemini 2.0 Flash, GPT-4o) and two agent architectures (ReAct, Reflexion) across four domains (scheduling, travel, customer support, e-commerce) over 1,280 episodes. Perturbations alone reduce success from 96.9% at ε=0 to 88.1% at ε=0.2. Rate limiting is the most damaging fault in ablations. ReAct is more robust than Reflexion under combined stress, and Gemini 2.0 Flash achieves comparable reliability to GPT-4o at much lower cost. ReliabilityBench provides a systematic framework for assessing production readiness of LLM agents.