ITG NowHome › AI glossary › Humanity's Last Exam
What is Humanity's Last Exam?
Humanity's Last Exam is a very hard test with expert questions that is built to challenge the best AI models.
- Category Benchmarks
- Also called HLE
- Updated 20 Oct 2026
Humanity's Last Exam, explained simply
Humanity's Last Exam, or HLE, is a test with hard questions from many fields. Experts wrote them to stump AI.
The score is the share answered correctly. Higher is better.
This matters because older tests got too easy. HLE gives researchers room to see progress.
The name is a bit dramatic. Passing it would not mean an AI can do every human job.
On ITG Now, the AI models page lists Humanity's Last Exam scores, with a link to the source.
Related terms
BenchmarkGPQA DiamondARC-AGIReasoning modelLeaderboard
Where you'll see it on ITG Now
- AI newsDeepSeek says free model update tops a coding test
- AI modelClaude Fable 5.1
- AI modelClaude Haiku 5.5
- AI modelClaude Opus 5.5
- AI modelDeepSeek-V4.1-Flash
- AI modelGemma 4 31B
Questions people ask about Humanity's Last Exam
Is a higher Humanity's Last Exam score better?
Yes. A higher score means more questions were answered correctly.
Why is it called the last exam?
The makers meant it to be the last school style test AI would need. The name is a bold claim, not a fact.
Does a high score mean AGI?
No. It is one test. It doesn't prove a model can do everything people can.