Humanity's Last Exam

Humanity's Last Exam Dataset

Humanity's Last Exam Humanity's Last Exam Paper Nature Arxiv Dataset load_dataset(" cais/hle ") GitHub HLE-Rolling Live Submission Dashboard Latest News [01/28/2026] : HLE is published on Nature (Nature 649, 1139–1146). [10/08/2025] : We release a dynamic fork version HLE-Rolling ( update logs ). To contribute, please send us an email at agibenchmark@safe.ai See full notes [04/03/2025] : HLE has been finalized with 2,500 questions. Questions flagged in the bug bounty program and searchable questions have been removed and replaced with other questions. [03/21/2025] : Bug Bounty closed. Thank yo

related reading

• Humanity's Last Machinehumanityslastmachine.com
• Elicit: AI for scientific researchelicit.org
• Parsed | Custom, interpretable AI systems that continuously learnparsed.com
• Elicit: AI for scientific researchelicit.com
• benchmarks.bio — Agentic AI benchmarks on messy, real-world biological databenchmarks.bio
• Dataset list - A list of the biggest machine learning datasetsdatasetlist.com
• ML Contestsmlcontests.com
• AI Model & API Providers Analysis | Artificial Analysisartificialanalysis.ai
• GitHub - open-compass/opencompass: OpenCompass is an LLM evaluation platform, supporting a wide range of models (Llama3, Mistral, InternLM2,GPT-4,LLaMa2, Qwen,GLM, Claude, etc) over 100+ datasets. · GitHubgithub.com
• AI Agent Benchmark for Real-World Professional Workflowsagents-last-exam.org
• Humanoid Atlas | Humanoid Robot Supply Chain Map, OEM Database & Industry Analysishumanoids.fyi
• AI Benchmark Leaderboards & Model Evals | BenchmarkListbenchmarklist.com