A benchmark of expert-level academic questions to assess AI capabilities

[EN] Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve more than 90% accuracy on popular benchmarks such as Measuring Massive Multitask Language Understanding(1), lim...

ver descrição completa

Detalhes bibliográficos
Autores: Phan, Long, Gatti, Alice, Li, Nathaniel, Khoja, Adam, Kim, Ryan, Ren, Richard, Hausenloy, Jason, Zhang, Oliver, Mazeika, Mantas, Hendrycks, Dan, Han, Ziwen, Hu, Josephina, Zhang, Hugh, Zhang, Chen Bo Calvin, Shaaban, Mohamed, Hernández-Orallo, José|||0000-0001-9746-7632
Formato: artículo
Fecha de publicación:2026
País:España
Recursos:Universitat Politècnica de València (UPV)
Repositorio:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Idioma:inglés
OAI Identifier:oai:dnet:riunet______::1bff9ebbc95c82aaca54cfc5b941466a
Acesso em linha:https://riunet.upv.es/handle/10251/233523
Access Level:acceso abierto
Palavra-chave:Large language models
Benchmark evaluation
Humanity s Last Exam
Multimodal benchmarks
AI capability assessment
Model calibration
Descrição
Resumo:[EN] Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve more than 90% accuracy on popular benchmarks such as Measuring Massive Multitask Language Understanding(1), limiting informed measurement of state-of-the-art LLM capabilities. Here, in response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be an expert-level closed-ended academic benchmark with broad subject coverage. HLE consists of 2,500 questions across dozens of subjects, including mathematics, humanities and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable but cannot be quickly answered by internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a marked gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai.