- Organizations
- New York University
New York University: AI benchmarks & research
NYU Center for Data Science and courant groups. Maintains GPQA (with Cohere and Anthropic co-authors), SQuALITY and NYU CTF Bench, and co-maintains LiveBench.
Maintained by New York University
8 benchmarks- Cybersecurity CTFs
Cybersecurity Capture the Flag (CTF) benchmark for evaluating LLMs in offensive security challenges. Contains diverse cybersecurity tasks including cryptography, web exploitation, binary analysis, and forensics to assess AI capabilities in cybersecurity problem-solving.
- Leader
GPT-5.3 Codex77.6%
- Models scored
- 3
- GPQA
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
- Leader
GPT-6 Astra96.0%
- Models scored
- 265
- GPQA Biology
Biology subset of GPQA, containing challenging multiple-choice questions written by domain experts in biology. These Google-proof questions require graduate-level knowledge and reasoning.
- Leader
o169.2%
- Models scored
- 1
- GPQA Chemistry
Chemistry subset of GPQA, containing challenging multiple-choice questions written by domain experts in chemistry. These Google-proof questions require graduate-level knowledge and reasoning.
- Leader
o164.7%
- Models scored
- 1
- GPQA Physics
Physics subset of GPQA, containing challenging multiple-choice questions written by domain experts in physics. These Google-proof questions require graduate-level knowledge and reasoning.
- Leader
o192.8%
- Models scored
- 1
- SQuALITY
SQuALITY (Summarization-format QUestion Answering with Long Input Texts, Yes!) is a long-document summarization dataset built by hiring highly-qualified contractors to read public-domain short stories (3000-6000 words) and write original summaries from scratch. Each document has five summaries: one overview and four question-focused summaries. Designed to address limitations in existing summarization datasets by providing high-quality, faithful summaries.
- Leader
Phi-3.5-mini-instruct24.3%
- Models scored
- 5
- LiveBench
Primary maintainer Abacus.AI
LiveBench is a challenging, contamination-limited LLM benchmark that addresses test set contamination by releasing new questions monthly based on recently-released datasets, arXiv papers, news articles, and IMDb movie synopses. It comprises tasks across math, coding, reasoning, language, instruction following, and data analysis with verifiable, objective ground-truth answers.
- Leader
o3-mini84.6%
- Models scored
- 38
- LiveBench 20241125
Primary maintainer Abacus.AI
LiveBench is a challenging, contamination-limited LLM benchmark that addresses test set contamination by releasing new questions monthly based on recently-released datasets, arXiv papers, news articles, and IMDb movie synopses. It comprises tasks across math, coding, reasoning, language, instruction following, and data analysis with verifiable, objective ground-truth answers.
- Leader
Qwen3 VL 235B A22B Thinking79.6%- Models scored
- 15