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

By Long Phan, Alice Gatti, Ziwen Han, Nathaniel Li, Josephina Hu, Hugh Zhang, Chen Bo Calvin Zhang, Sean Shi, Michael Choi, Anirudh Agrawal and 668 more. Nature, 2025

349 citations recorded by the source corpus.

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 Understanding1, 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. Humanity’s Last Exam, a multi-modal benchmark at the frontier of human knowledge, is designed to be an expert-level closed-ended academic benchmark with broad subject coverage.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/a5524d085ac586d531021dcb1ec156eaf942b109

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Kimi K2: Open Agentic Intelligence (2025, positive): “In reasoning, we include a wide range of mathematical, science and logical tasks: AIME 2024/2025, MATH-500, HMMT 2025, CNMO 2024, PolyMath-en, ZebraLogic [44], AutoLogi [92], GPQA-Diamond [62], SuperGPQA [14], and Humanity’s Last Exam (Text-Only) [57].”
LLMs Get Lost In Multi-Turn Conversation (2025, positive): “…effort has been put in improving LLM aptitude : demonstrating that LLMs can accomplish tasks of increasing intellectual complexity, with recent results showing LLMs can compete in mathematics Olympiads, or solve Ph.D.-level technical questions in a benchmark aptly named Humanity’s Last Exam [62].”
WebThinker: Empowering Large Reasoning Models with Deep Research Capability (2025, neutral): “Use <|begin_write_section|> section name\ncontents to write <|end_write_section|> to call the system to write a section in the article -”
BrowseComp: A Simple Yet Challenging Benchmark for Browsing Agents (2025, positive): “We use the same grading prompt as Humanity’s Last Exam (Phan et al., 2025).”
gpt-oss-120b&gpt-oss-20b Model Card (2025, neutral): “We evaluate on: • Reasoning and factuality : AIME, GPQA [24], MMLU [25], and HLE [26].”
GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models (2025, neutral): “5-Air on seven benchmarks, including MMLU-Pro [43], AIME 24, MATH 500 [14], SciCode [36], GPQA [30], Humanity’s Last Exam (HLE) [28], and LiveCodeBench (LCB) [19] 2 .”
R1-Onevision: Advancing Generalized Multimodal Reasoning Through Cross-Modal Formalization (2025, neutral): “Although some more challenging datasets, like HumanEval-V [37], Humanity’s Last Exam [20], and ZeroBench [23], aim to assess complex diagram understanding or broader knowledge proficiency, these benchmarks remain specialized, covering only limited aspects of reasoning.”
ResearchGPT: Benchmarking and Training LLMs for End-to-End Computer Science Research Workflows (2025, unknown): “, GPQA (Rein et al., 2024), SuperGPQA (Du et al., 2025), and Humanity’s Last Exam (HLE) (Phan et al., 2025) test”