# Do Large Language Model Benchmarks Test Reliability?

arXiv: https://arxiv.org/abs/2502.03461 · HTML: https://mlsanity.com/paper/2502.03461/

## Key facts

- Cited by 54 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Feb 2025.

## Built on by (most cited first, 30 of 54)

- [Kimi K2: Open Agentic Intelligence](https://mlsanity.com/paper/2507.20534/) (2025)
- [The Illusion of Diminishing Returns: Measuring Long Horizon Execution in LLMs](https://mlsanity.com/paper/2509.09677/) (2025)
- [Rethinking Reflection in Pre-Training](https://mlsanity.com/paper/2504.04022/) (2025)
- [Prismatic Synthesis: Gradient-based Data Diversification Boosts Generalization in LLM Reasoning](https://mlsanity.com/paper/2505.20161/) (2025)
- [Adaptive Testing for LLM Evaluation: A Psychometric Alternative to Static Benchmarks](https://mlsanity.com/paper/2511.04689/) (2025)
- [Signal and Noise: A Framework for Reducing Uncertainty in Language Model Evaluation](https://mlsanity.com/paper/2508.13144/) (2025)
- [Antidistillation Sampling](https://mlsanity.com/paper/2504.13146/) (2025)
- [Noisy but Valid: Robust Statistical Evaluation of LLMs with Imperfect Judges](https://mlsanity.com/paper/2601.20913/) (2026)
- [Evaluating LLM Metrics Through Real-World Capabilities](https://mlsanity.com/paper/2505.08253/) (2025)
- [WUSH: Near-Optimal Adaptive Transforms for LLM Quantization](https://mlsanity.com/paper/2512.00956/) (2025)
- [Verifying the Verifiers: Unveiling Pitfalls and Potentials in Fact Verifiers](https://mlsanity.com/paper/2506.13342/) (2025)
- [Simple Policy Gradients for Reasoning with Diffusion Language Models](https://mlsanity.com/paper/2510.04019/) (2025)
- [Automated Benchmark Auditing for AI Agents and Large Language Models](https://mlsanity.com/paper/2605.26079/) (2026)
- [Who Benchmarks the Benchmarks? Towards Comprehensive Evaluation of Commonsense Reasoning Benchmarks](https://mlsanity.com/paper/2504.07825/) (2025)
- [Fantastic Bugs and Where to Find Them in AI Benchmarks](https://mlsanity.com/paper/2511.16842/) (2025)
- [Benchmarks Saturate When The Model Gets Smarter Than The Judge](https://mlsanity.com/paper/2601.19532/) (2026)
- [Min-p, Max Exaggeration: A Critical Analysis of Min-p Sampling in Language Models](https://mlsanity.com/paper/2506.13681/) (2025)
- [Scaling Graph Chain-of-Thought Reasoning: A Multi-Agent Framework with Efficient LLM Serving](https://mlsanity.com/paper/2511.01633/) (2025)
- [Benchmarking Misuse Mitigation Against Covert Adversaries](https://mlsanity.com/paper/2506.06414/) (2025)
- [CapBencher: Give Your LLM Benchmark a Built-in Alarm for Test-Set Overfitting](https://mlsanity.com/paper/2505.18102/) (2025)
- [Position: Machine Learning Conferences Should Establish a "Refutations and Critiques" Track](https://mlsanity.com/paper/2506.19882/) (2025)
- [DESIGNER: Design-Logic-Guided Multidisciplinary Data Synthesis for LLM Reasoning](https://mlsanity.com/paper/2508.12726/) (2025)
- [WildSci: Advancing Scientific Reasoning from In-the-Wild Literature](https://mlsanity.com/paper/2601.05567/) (2026)
- [Protecting Language Models Against Unauthorized Distillation through Trace Rewriting](https://mlsanity.com/paper/2602.15143/) (2026)
- [Beyond the Mean: Within-Model Reliable Change Detection for LLM Evaluation](https://mlsanity.com/paper/2604.27405/) (2026)
- [Learning a Continue-Thinking Token for Enhanced Test-Time Scaling](https://mlsanity.com/paper/2506.11274/) (2025)
- [Weight Decay Improves Language Model Plasticity](https://mlsanity.com/paper/2602.11137/) (2026)
- [Panini: Continual Learning in Token Space via Structured Memory](https://mlsanity.com/paper/2602.15156/) (2026)
- [Domain Specific Benchmarks for Evaluating Multimodal Large Language Models](https://mlsanity.com/paper/2506.12958/) (2025)
- [Rethinking Math Reasoning Evaluation: A Robust LLM-as-a-Judge Framework Beyond Symbolic Rigidity](https://mlsanity.com/paper/2604.22597/) (2026)

## Abstract (arXiv)

When deploying large language models (LLMs), it is important to ensure that these models are not only capable, but also reliable. Many benchmarks have been created to track LLMs' growing capabilities, however there has been no similar focus on measuring their reliability. To understand the potential ramifications of this gap, we investigate how well current benchmarks quantify model reliability. We find that pervasive label errors can compromise these evaluations, obscuring lingering model failures and hiding unreliable behavior. Motivated by this gap in the evaluation of reliability, we then propose the concept of so-called platinum benchmarks, i.e., benchmarks carefully curated to minimize label errors and ambiguity. As a first attempt at constructing such benchmarks, we revise examples from fifteen existing popular benchmarks. We evaluate a wide range of models on these platinum benchmarks and find that, indeed, frontier LLMs still exhibit failures on simple tasks such as elementary-level math word problems. Analyzing these failures further reveals previously unidentified patterns of problems on which frontier models consistently struggle. We provide code at https://github.com/MadryLab/platinum-benchmarks