aiqbench analysis
This research proposes a new metric called "accessibility" to evaluate large language models by measuring success rates across multiple equivalent prompt formulations rather than relying on single fixed wordings. The study demonstrates that this approach uncovers critical trade-offs between broad capability coverage and consistent reliability, revealing significant divergences in model performance that standard accuracy scores miss. By distinguishing between models that succeed occasionally versus those that are dependable, this work offers a more robust framework for assessing the real-world utility of AI systems.
Revealing what benchmark scores hide: Jointly measuring the breadth and reliability of LLM capabilities using accessibility profiles
From arXiv
Full textarXiv:2610.24957v1 Announce Type: cross Abstract: Language model benchmarks assign each question a single outcome from one fixed wording, even though the same underlying request can be expressed in many equivalent ways. We argue that the quantity of interest should therefore not be correctness on one wording, but the accessibility of an inquiry, which we define as the probability that a model succeeds across meaning-preserving formulations. The distribution of accessibility across inquiries reveals a structure that accuracy and existing best-, worst-, or average-case prompt scores collapse: how broadly a model's capabilities extend across inquiries and how dependably those capabilities can be elicited.
Two models can have the same average performance while one succeeds across more inquiries unreliably and the other succeeds across fewer inquiries but more consistently, leading to opposite preferences depending on whether the task allows multiple retries until one success or when repeated consistent success is required. Across six benchmarks and more than twenty models, this tradeoff is common: for about half of the model pairs with nearly equal mean accessibility, which model looks better depends on whether one success across five formulations is enough or all five must succeed, quantities we call Access@5 and Reliable@5, respectively.
For some pairs the two measures diverge significantly, by up to 10 points in opposite directions. Accessibility also reveals that scaling, reasoning, and prompt optimization can redistribute capability across inquiries in ways aggregate accuracy obscures, sometimes broadening the set of inquiries a model can succeed on while making success on them less repeatable. Accessibility turns formulation sensitivity from a nuisance into a measurable property of capability, distinguishing what a model can sometimes succeed on from what users can actually depend on.
Abstract:Language model benchmarks assign each question a single outcome from one fixed wording, even though the same underlying request can be expressed in many equivalent ways. We argue that the quantity of interest should therefore not be correctness on one wording, but the accessibility of an inquiry, which we define as the probability that a model succeeds across meaning-preserving formulations. The distribution of accessibility across inquiries reveals a structure that accuracy and existing best-, worst-, or average-case prompt scores collapse: how broadly a model's capabilities extend across inquiries and how dependably those capabilities can be elicited.
Two models can have the same average performance while one succeeds across more inquiries unreliably and the other succeeds across fewer inquiries but more consistently, leading to opposite preferences depending on whether the task allows multiple retries until one success or when repeated consistent success is required. Across six benchmarks and more than twenty models, this tradeoff is common: for about half of the model pairs with nearly equal mean accessibility, which model looks better depends on whether one success across five formulations is enough or all five must succeed, quantities we call Access@5 and Reliable@5, respectively.
For some pairs the two measures diverge significantly, by up to 10 points in opposite directions. Accessibility also reveals that scaling, reasoning, and prompt optimization can redistribute capability across inquiries in ways aggregate accuracy obscures, sometimes broadening the set of inquiries a model can succeed on while making success on them less repeatable. Accessibility turns formulation sensitivity from a nuisance into a measurable property of capability, distinguishing what a model can sometimes succeed on from what users can actually depend on.
26 pages total, 10 pages main text, 18 figures. Code and data: this https URL
Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
arXiv:2610.24957 [cs.AI]
(or arXiv:2610.24957v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2610.24957 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 388 words
Related coverage
A Multi-Tiered AI Accessibility Model (MAIAM): Proportionate Access and Capability Limits for Comprehensive Frontier AI Safety
arXiv · Oct 17, 2026
Learning to Access Computation: Accessibility Plasticity as a Principle of Adaptive Intelligence
arXiv · Jul 28, 2026

Regulate the Deployment, Not Just the Model
LessWrong · Oct 31, 2026
ReSight: Task-Driven Self-Improving Memory for Embodied Agents
arXiv · Oct 27, 2026