# BenchGuard: Who Guards the Benchmarks? Automated Auditing of LLM Agent Benchmarks

arXiv: https://arxiv.org/abs/2604.24955 · HTML: https://mlsanity.com/paper/2604.24955/

## Key facts

- Cited by 15 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 21 of its 32 references that are in the map.
- First posted on arXiv 27 Apr 2026.

## Placed in

- LLM agents and tool use / [General LLM agent frameworks and benchmarks](https://mlsanity.com/agents/general-llm-agent-frameworks-and-benchmarks/)

## Builds on (most cited first, 21 of 21)

- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [Is Your Code Generated by ChatGPT Really Correct? Rigorous Evaluation of Large Language Models for Code Generation](https://mlsanity.com/paper/2305.01210/) (2023)
- [Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces](https://mlsanity.com/paper/2601.11868/) (2026)
- [From Generation to Judgment: Opportunities and Challenges of LLM-as-a-judge](https://mlsanity.com/paper/2411.16594/) (2024)
- [Humanity's Last Exam](https://mlsanity.com/paper/2501.14249/) (2025)
- [Prometheus 2: An Open Source Language Model Specialized in Evaluating Other Language Models](https://mlsanity.com/paper/2405.01535/) (2024)
- [Pervasive Label Errors in Test Sets Destabilize Machine Learning Benchmarks](https://mlsanity.com/paper/2103.14749/) (2021)
- [ScienceAgentBench: Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discovery](https://mlsanity.com/paper/2410.05080/) (2024)
- [What Will it Take to Fix Benchmarking in Natural Language Understanding?](https://mlsanity.com/paper/2104.02145/) (2021)
- [Are We Done with MMLU?](https://mlsanity.com/paper/2406.04127/) (2024)
- [Towards End-to-End Automation of AI Research](https://mlsanity.com/paper/2606.15497/) (2026)
- [LLM Critics Help Catch LLM Bugs](https://mlsanity.com/paper/2407.00215/) (2024)
- [BixBench: a Comprehensive Benchmark for LLM-based Agents in Computational Biology](https://mlsanity.com/paper/2503.00096/) (2025)
- [Are "Solved Issues" in SWE-bench Really Solved Correctly? An Empirical Study](https://mlsanity.com/paper/2503.15223/) (2025)
- [UTBoost: Rigorous Evaluation of Coding Agents on SWE-Bench](https://mlsanity.com/paper/2506.09289/) (2025)
- [HLE-Verified: A Systematic Verification and Structured Revision of Humanity's Last Exam](https://mlsanity.com/paper/2602.13964/) (2026)
- [Benchmarks Saturate When The Model Gets Smarter Than The Judge](https://mlsanity.com/paper/2601.19532/) (2026)
- [The Fault in our Stars: Quality Assessment of Code Generation Benchmarks](https://mlsanity.com/paper/2404.10155/) (2024)
- [ELT-Bench-Verified: Benchmark Quality Issues Underestimate AI Agent Capabilities](https://mlsanity.com/paper/2603.29399/) (2026)

## Built on by (most cited first, 15 of 15)

- [Do Androids Dream of Breaking the Game? Systematically Auditing AI Agent Benchmarks with BenchJack](https://mlsanity.com/paper/2605.12673/) (2026)
- [Automated Benchmark Auditing for AI Agents and Large Language Models](https://mlsanity.com/paper/2605.26079/) (2026)
- [Auditing Harness Tampering in Self-Improving Agents](https://mlsanity.com/paper/2609.00069/) (2026)
- [Cost-Utility Alignment in LLM Agent Trajectories:Profiling,Attribution,Diagnosis,Adaptation,and Evaluation](https://mlsanity.com/paper/2608.26195/) (2026)
- [Do Agent Benchmarks Do What They Say? An Executable-Contract Audit of Tool-Using Agent Environments](https://mlsanity.com/paper/2609.37315/) (2026)
- [D3-Gym: Constructing Real-World Verifiable Environments for Data-Driven Discovery](https://mlsanity.com/paper/2604.27977/) (2026)
- [Automated Transcript Analysis for Detecting Flaws in Agentic Benchmarks](https://mlsanity.com/paper/2607.27518/) (2026)
- [Maintaining Benchmarks Against Increasingly Capable Agents: Detection and Remediation of Unearned Passes](https://mlsanity.com/paper/2609.34262/) (2026)
- [AuditRepairBench: A Paired-Execution Trace Corpus for Evaluator-Channel Ranking Instability in Agent Repair](https://mlsanity.com/paper/2605.04624/) (2026)
- [What Makes Software Issue Resolution Tasks Difficult for Agents?](https://mlsanity.com/paper/2608.18280/) (2026)
- [SimVerity: When Does Simulated Agent Success Survive Physical Deployment?](https://mlsanity.com/paper/2608.25067/) (2026)
- [MLIP Detective: Active Failure Mode Discovery Beyond Benchmark Scores for Machine-Learning Interatomic Potentials](https://mlsanity.com/paper/2609.08399/) (2026)
- [What Did the Agent Actually Do? Evidence-Grounded Oversight for Long-Horizon Agents](https://mlsanity.com/paper/2610.06406/) (2026)
- [A Trust Layer for Agent Evaluation](https://mlsanity.com/paper/2610.07274/) (2026)
- [TestJack: Should you trust the results in coding benchmarks? Agentic Coding Benchmarks Auditing via Evaluator Evolution](https://mlsanity.com/paper/2610.10619/) (2026)

## Abstract (arXiv)

As benchmarks grow in complexity, many apparent agent failures are not failures of the agent at all---they are failures of the benchmark itself: broken specifications, implicit assumptions, and rigid evaluation scripts that penalize valid alternative approaches. We propose employing frontier LLMs as systematic auditors of evaluation infrastructure, and realize this vision through BenchGuard, the first framework explicitly designed for joint cross-artifact auditing of execution-based agent benchmarks. BenchGuard cross-verifies all benchmark artifacts via structured LLM protocols, optionally incorporating agent solutions or execution traces as additional diagnostic evidence. Deployed on two prominent scientific benchmarks, BenchGuard identified 12 author-confirmed issues in ScienceAgentBench---including fatal errors rendering tasks unsolvable---and exactly matched 83.3% of expert-identified issues on the BIXBench Verified-50 subset, catching defects that prior human review missed entirely. A full audit of 50 complex bioinformatics tasks costs under USD 15, making automated benchmark auditing a practical and valuable complement to human review. A preliminary native-format audit of ProgramBench further demonstrates cross-format applicability. These findings point toward AI-assisted benchmark development, where frontier models serve not only as subjects of evaluation but as active participants in validating the evaluation infrastructure itself.