Jump to content

Humanity's Last Exam

From Wikipedia, the free encyclopedia
(Redirected from HLE (Humanity's Last Exam))

Humanity's Last Exam (HLE) is a language model benchmark consisting of 2,500 questions across a broad range of subjects. It was created jointly by the Center for AI Safety and Scale AI.

Creation

[edit]

Stanford HAI's AI Index 2025 Annual Report cites Humanity's Last Exam as one of the "more challenging benchmarks" developed in response to the popular AI benchmarks having reached "saturation".[1] The test has been described as the brainchild of Dan Hendrycks, a machine learning researcher and the director of the Center for AI Safety, who stated that he was inspired to create the test after a conversation with Elon Musk, who thought the existing language model benchmarks, such as the MMLU, were too easy. Hendrycks worked with Scale AI to compile the questions.[2] The questions were crowdsourced from subject matter experts from various institutions across the world.[3][4] The questions were first filtered by the leading AI models; if the models failed to answer the question or did worse than random guessing on the multiple-choice questions, they were reviewed by human experts in two rounds and approved for inclusion in the dataset. The submitters of the top-rated questions were given prize money from a pool of 500,000 U.S. dollars—$5000 for each of the top 50 questions and $500 for the next 500. After the initial release, a "community feedback bug bounty program" was opened to "identify and remove major errors in the dataset".[4]

Composition

[edit]

The benchmark consists of 2,500 questions in the publicly released set. The paper classifies the questions into the following broad subjects: mathematics (41%), physics (9%), biology/medicine (11%), humanities/social science (9%), computer science/artificial intelligence (10%), engineering (4%), chemistry (7%), and other (9%). Around 14% of the questions require the ability to understand both text and images, i.e., multi-modality. 24% of the questions are multiple-choice; the rest are short-answer, exact-match questions. A private set is also maintained to test for benchmark overfitting.[4]

An example question:[2]

Hummingbirds within Apodiformes uniquely have a bilaterally paired oval bone, a sesamoid embedded in the caudolateral portion of the expanded, cruciate aponeurosis of insertion of m. depressor caudae. How many paired tendons are supported by this sesamoid bone? Answer with a number.

Results

[edit]
Performance of various models on the benchmark
Organization Model Accuracy (%) ↑ Calibration Error (%) ↓
OpenAI o3 (high) 20.32 34
Google DeepMind Gemini 2.5 Pro Experimental 18.16 71
OpenAI o3-mini (high) 13.37[a] 80
DeepSeek DeepSeek-R1 8.54[a] 73
Anthropic Claude 3.7 Sonnet (16K) 8.04 80
Source: Center for AI Safety and Scale AI. 17 April 2025.

Notes

[edit]

Footnotes

[edit]
  1. ^ a b o3-mini (high) and DeepSeek-R1 are not multimodal models and were evaluated only on the text-only subset.

References

[edit]
  1. ^ Maslej, Nestor; et al. (April 2025). The AI Index 2025 Annual Report (PDF) (Report). Institute for Human-Centered AI. pp. 141–142.
  2. ^ a b Roose, Kevin (23 January 2025). "When A.I. Passes This Test, Look Out". New York Times. Archived from the original on 29 January 2025. Retrieved 24 January 2025.
  3. ^ Dastin, Jeffrey; Paul, Katie (16 September 2024). "AI experts ready 'Humanity's Last Exam' to stump powerful tech". Reuters. Archived from the original on 8 April 2025. Retrieved 24 January 2025.
  4. ^ a b c Phan, Long; et al. (2025). "Humanity's Last Exam". arXiv:2501.14249 [cs.LG].
[edit]