PaperlayerView exact revision on arXiv

Source-linked AI summary

Humanity's Last Exam

Long Phan, Alice Gatti, Ziwen Han, Nathaniel Li, Josephina Hu, Hugh Zhang, Chen Bo Calvin Zhang, Mohamed Shaaban, John Ling, Sean Shi, Michael Choi, Anish Agrawal, Arnav Chopra, Adam Khoja, Ryan Kim, Richard Ren, Jason Hausenloy, Oliver Zhang, Mantas Mazeika, Dmitry Dodonov, Tung Nguyen, Jaeho Lee, Daron Anderson, Mikhail Doroshenko, Alun Cennyth Stokes, Mobeen Mahmood, Oleksandr Pokutnyi, Oleg Iskra, Jessica P. Wang, John-Clark Levin, Mstyslav Kazakov, Fiona Feng, Steven Y. Feng, Haoran Zhao, Michael Yu, Varun Gangal, Chelsea Zou, Zihan Wang, Serguei Popov, Robert Gerbicz, Geoff Galgon, Johannes Schmitt, Will Yeadon, Yongki Lee, Scott Sauers, Alvaro Sanchez, Fabian Giska, Marc Roth, Søren Riis, Saiteja Utpala, Noah Burns, Gashaw M. Goshu, Mohinder Maheshbhai Naiya, Chidozie Agu, Zachary Giboney, Antrell Cheatom, Francesco Fournier-Facio, Sarah-Jane Crowson, Lennart Finke, Zerui Cheng, Jennifer Zampese, Ryan G. Hoerr, Mark Nandor, Hyunwoo Park, Tim Gehrunger, Jiaqi Cai, Ben McCarty, Alexis C Garretson, Edwin Taylor, Damien Sileo, Qiuyu Ren, Usman Qazi, Lianghui Li, Jungbae Nam, John B. Wydallis, Pavel Arkhipov, Jack Wei Lun Shi, Aras Bacho, Chris G. Willcocks, Hangrui Cao, Sumeet Motwani, Emily de Oliveira Santos, Johannes Veith, Edward Vendrow, Doru Cojoc, Kengo Zenitani, Joshua Robinson, Longke Tang, Yuqi Li, Joshua Vendrow, Natanael Wildner Fraga, Vladyslav Kuchkin, Andrey Pupasov Maksimov, Pierre Marion, Denis Efremov, Jayson Lynch, Kaiqu Liang, Aleksandar Mikov, Andrew Gritsevskiy, Julien Guillod, Gözdenur Demir, Dakotah Martinez, Ben Pageler, Kevin Zhou, Saeed Soori, Ori Press, Henry Tang, Paolo Rissone, Sean R. Green, Lina Brüssel, Moon Twayana, Aymeric Dieuleveut, Joseph Marvin Imperial, Ameya Prabhu, Jinzhou Yang, Nick Crispino, Arun Rao, Dimitri Zvonkine, Gabriel Loiseau, Mikhail Kalinin, Marco Lukas, Ciprian Manolescu, Nate Stambaugh, Subrata Mishra, Tad Hogg, Carlo Bosio, Brian P Coppola, Julian Salazar, Jaehyeok Jin, Rafael Sayous, Stefan Ivanov, Philippe Schwaller, Shaipranesh Senthilkuma, Andres M Bran, Andres Algaba, Kelsey Van den Houte, Lynn Van Der Sypt, Brecht Verbeken, David Noever, Alexei Kopylov, Benjamin Myklebust, Bikun Li, Lisa Schut, Evgenii Zheltonozhskii, Qiaochu Yuan, Derek Lim, Richard Stanley, Tong Yang, John Maar, Julian Wykowski, Martí Oller, Anmol Sahu, Cesare Giulio Ardito, Yuzheng Hu, Ariel Ghislain Kemogne Kamdoum, Alvin Jin, Tobias Garcia Vilchis, Yuexuan Zu, Martin Lackner, James Koppel, Gongbo Sun, Daniil S. Antonenko, Steffi Chern, Bingchen Zhao, Pierrot Arsene, Joseph M Cavanagh, Daofeng Li, Jiawei Shen, Donato Crisostomi, Wenjin Zhang, Ali Dehghan, Sergey Ivanov, David Perrella, Nurdin Kaparov, Allen Zang, Ilia Sucholutsky, Arina Kharlamova, Daniil Orel, Vladislav Poritski, Shalev Ben-David, Zachary Berger, Parker Whitfill, Michael Foster, Daniel Munro, Linh Ho, Shankar Sivarajan, Dan Bar Hava, Aleksey Kuchkin, David Holmes, Alexandra Rodriguez-Romero, Frank Sommerhage, Anji Zhang, Richard Moat, Keith Schneider, Zakayo Kazibwe, Don Clarke, Dae Hyun Kim, Felipe Meneguitti Dias, Sara Fish, Veit Elser, Tobias Kreiman, Victor Efren Guadarrama Vilchis, Immo Klose, Ujjwala Anantheswaran, Adam Zweiger, Kaivalya Rawal, Jeffery Li, Jeremy Nguyen, Nicolas Daans, Haline Heidinger, Maksim Radionov, Václav Rozhoň, Vincent Ginis, Christian Stump, Niv Cohen, Rafał Poświata, Josef Tkadlec, Alan Goldfarb, Chenguang Wang, Piotr Padlewski, Stanislaw Barzowski, Kyle Montgomery, Ryan Stendall, Jamie Tucker-Foltz, Jack Stade, T. Ryan Rogers, Tom Goertzen, Declan Grabb, Abhishek Shukla, Alan Givré, John Arnold Ambay, Archan Sen, Muhammad Fayez Aziz, Mark H Inlow, Hao He, Ling Zhang, Younesse Kaddar, Ivar Ängquist, Yanxu Chen, Harrison K Wang, Kalyan Ramakrishnan, Elliott Thornley, Antonio Terpin, Hailey Schoelkopf, Eric Zheng, Avishy Carmi, Ethan D. L. Brown, Kelin Zhu, Max Bartolo, Richard Wheeler, Martin Stehberger, Peter Bradshaw, JP Heimonen, Kaustubh Sridhar, Ido Akov, Jennifer Sandlin, Yury Makarychev, Joanna Tam, Hieu Hoang, David M. Cunningham, Vladimir Goryachev, Demosthenes Patramanis, Michael Krause, Andrew Redenti, David Aldous, Jesyin Lai, Shannon Coleman, Jiangnan Xu, Sangwon Lee, Ilias Magoulas, Sandy Zhao, Ning Tang, Michael K. Cohen, Orr Paradise, Jan Hendrik Kirchner, Maksym Ovchynnikov, Jason O. Matos, Adithya Shenoy, Michael Wang, Yuzhou Nie, Anna Sztyber-Betley, Paolo Faraboschi, Robin Riblet, Jonathan Crozier, Shiv Halasyamani, Shreyas Verma, Prashant Joshi, Eli Meril, Ziqiao Ma, Jérémy Andréoletti, Raghav Singhal, Jacob Platnick, Volodymyr Nevirkovets, Luke Basler, Alexander Ivanov, Seri Khoury, Nils Gustafsson, Marco Piccardo, Hamid Mostaghimi, Qijia Chen, Virendra Singh, Tran Quoc Khánh, Paul Rosu, Hannah Szlyk, Zachary Brown, Himanshu Narayan, Aline Menezes, Jonathan Roberts, William Alley, Kunyang Sun, Arkil Patel, Max Lamparth, Anka Reuel, Linwei Xin, Hanmeng Xu, Jacob Loader, Freddie Martin, Zixuan Wang, Andrea Achilleos, Thomas Preu, Tomek Korbak, Ida Bosio, Fereshteh Kazemi, Ziye Chen, Biró Bálint, Eve J. Y. Lo, Jiaqi Wang, Maria Inês S. Nunes, Jeremiah Milbauer, M Saiful Bari, Zihao Wang, Behzad Ansarinejad, Yewen Sun, Stephane Durand, Hossam Elgnainy, Guillaume Douville, Daniel Tordera, George Balabanian, Hew Wolff, Lynna Kvistad, Hsiaoyun Milliron, Ahmad Sakor, Murat Eron, Andrew Favre D. O., Shailesh Shah, Xiaoxiang Zhou, Firuz Kamalov, Sherwin Abdoli, Tim Santens, Shaul Barkan, Allison Tee, Robin Zhang, Alessandro Tomasiello, G. Bruno De Luca, Shi-Zhuo Looi, Vinh-Kha Le, Noam Kolt, Jiayi Pan, Emma Rodman, Jacob Drori, Carl J Fossum, Niklas Muennighoff, Milind Jagota, Ronak Pradeep, Honglu Fan, Jonathan Eicher, Michael Chen, Kushal Thaman, William Merrill, Moritz Firsching, Carter Harris, Stefan Ciobâcă, Jason Gross, Rohan Pandey, Ilya Gusev, Adam Jones, Shashank Agnihotri, Pavel Zhelnov, Mohammadreza Mofayezi, Alexander Piperski, David K. Zhang, Kostiantyn Dobarskyi, Roman Leventov, Ignat Soroko, Joshua Duersch, Vage Taamazyan, Andrew Ho, Wenjie Ma, William Held, Ruicheng Xian, Armel Randy Zebaze, Mohanad Mohamed, Julian Noah Leser, Michelle X Yuan, Laila Yacar, Johannes Lengler, Katarzyna Olszewska, Claudio Di Fratta, Edson Oliveira, Joseph W. Jackson, Andy Zou, Muthu Chidambaram, Timothy Manik, Hector Haffenden, Dashiell Stander, Ali Dasouqi, Alexander Shen, Bita Golshani, David Stap, Egor Kretov, Mikalai Uzhou, Alina Borisovna Zhidkovskaya, Nick Winter, Miguel Orbegozo Rodriguez, Robert Lauff, Dustin Wehr, Colin Tang, Zaki Hossain, Shaun Phillips, Fortuna Samuele, Fredrik Ekström, Angela Hammon, Oam Patel, Faraz Farhidi, George Medley, Forough Mohammadzadeh, Madellene Peñaflor, Haile Kassahun, Alena Friedrich, Rayner Hernandez Perez, Daniel Pyda, Taom Sakal, Omkar Dhamane, Ali Khajegili Mirabadi, Eric Hallman, Kenchi Okutsu, Mike Battaglia, Mohammad Maghsoudimehrabani, Alon Amit, Dave Hulbert, Roberto Pereira, Simon Weber, Handoko, Anton Peristyy, Stephen Malina, Mustafa Mehkary, Rami Aly, Frank Reidegeld, Anna-Katharina Dick, Cary Friday, Mukhwinder Singh, Hassan Shapourian, Wanyoung Kim, Mariana Costa, Hubeyb Gurdogan, Harsh Kumar, Chiara Ceconello, Chao Zhuang, Haon Park, Micah Carroll, Andrew R. Tawfeek, Stefan Steinerberger, Daattavya Aggarwal, Michael Kirchhof, Linjie Dai, Evan Kim, Johan Ferret, Jainam Shah, Yuzhou Wang, Minghao Yan, Krzysztof Burdzy, Lixin Zhang, Antonio Franca, Diana T. Pham, Kang Yong Loh, Joshua Robinson, Abram Jackson, Paolo Giordano, Philipp Petersen, Adrian Cosma, Jesus Colino, Colin White, Jacob Votava, Vladimir Vinnikov, Ethan Delaney, Petr Spelda, Vit Stritecky, Syed M. Shahid, Jean-Christophe Mourrat, Lavr Vetoshkin, Koen Sponselee, Renas Bacho, Zheng-Xin Yong, Florencia de la Rosa, Nathan Cho, Xiuyu Li, Guillaume Malod, Orion Weller, Guglielmo Albani, Leon Lang, Julien Laurendeau, Dmitry Kazakov, Fatimah Adesanya, Julien Portier, Lawrence Hollom, Victor Souza, Yuchen Anna Zhou, Julien Degorre, Yiğit Yalın, Gbenga Daniel Obikoya, Rai, Filippo Bigi, M. C. Boscá, Oleg Shumar, Kaniuar Bacho, Gabriel Recchia, Mara Popescu, Nikita Shulga, Ngefor Mildred Tanwie, Thomas C. H. Lux, Ben Rank, Colin Ni, Matthew Brooks, Alesia Yakimchyk, Huanxu, Liu, Stefano Cavalleri, Olle Häggström, Emil Verkama, Joshua Newbould, Hans Gundlach, Leonor Brito-Santana, Brian Amaro, Vivek Vajipey, Rynaa Grover, Ting Wang, Yosi Kratish, Wen-Ding Li, Sivakanth Gopi, Andrea Caciolai, Christian Schroeder de Witt, Pablo Hernández-Cámara, Emanuele Rodolà, Jules Robins, Dominic Williamson, Vincent Cheng, Brad Raynor, Hao Qi, Ben Segev, Jingxuan Fan, Sarah Martinson, Erik Y. Wang, Kaylie Hausknecht, Michael P. Brenner, Mao Mao, Christoph Demian, Peyman Kassani, Xinyu Zhang, David Avagian, Eshawn Jessica Scipio, Alon Ragoler, Justin Tan, Blake Sims, Rebeka Plecnik, Aaron Kirtland, Omer Faruk Bodur, D. P. Shinde, Yan Carlos Leyva Labrador, Zahra Adoul, Mohamed Zekry, Ali Karakoc, Tania C. B. Santos, Samir Shamseldeen, Loukmane Karim, Anna Liakhovitskaia, Nate Resman, Nicholas Farina, Juan Carlos Gonzalez, Gabe Maayan, Earth Anderson, Rodrigo De Oliveira Pena, Elizabeth Kelley, Hodjat Mariji, Rasoul Pouriamanesh, Wentao Wu, Ross Finocchio, Ismail Alarab, Joshua Cole, Danyelle Ferreira, Bryan Johnson, Mohammad Safdari, Liangti Dai, Siriphan Arthornthurasuk, Isaac C. McAlister, Alejandro José Moyano, Alexey Pronin, Jing Fan, Angel Ramirez-Trinidad, Yana Malysheva, Daphiny Pottmaier, Omid Taheri, Stanley Stepanic, Samuel Perry, Luke Askew, Raúl Adrián Huerta Rodríguez, Ali M. R. Minissi, Ricardo Lorena, Krishnamurthy Iyer, Arshad Anil Fasiludeen, Ronald Clark, Josh Ducey, Matheus Piza, Maja Somrak, Eric Vergo, Juehang Qin, Benjámin Borbás, Eric Chu, Jack Lindsey, Antoine Jallon, I. M. J. McInnis, Evan Chen, Avi Semler, Luk Gloor, Tej Shah, Marc Carauleanu, Pascal Lauer, Tran Đuc Huy, Hossein Shahrtash, Emilien Duc, Lukas Lewark, Assaf Brown, Samuel Albanie, Brian Weber, Warren S. Vaz, Pierre Clavier, Yiyang Fan, Gabriel Poesia Reis e Silva, Long, Lian, Marcus Abramovitch, Xi Jiang, Sandra Mendoza, Murat Islam, Juan Gonzalez, Vasilios Mavroudis, Justin Xu, Pawan Kumar, Laxman Prasad Goswami, Daniel Bugas, Nasser Heydari, Ferenc Jeanplong, Thorben Jansen, Antonella Pinto, Archimedes Apronti, Abdallah Galal, Ng Ze-An, Ankit Singh, Tong Jiang, Joan of Arc Xavier, Kanu Priya Agarwal, Mohammed Berkani, Gang Zhang, Zhehang Du, Benedito Alves de Oliveira Junior, Dmitry Malishev, Nicolas Remy, Taylor D. Hartman, Tim Tarver, Stephen Mensah, Gautier Abou Loume, Wiktor Morak, Farzad Habibi, Sarah Hoback, Will Cai, Javier Gimenez, Roselynn Grace Montecillo, Jakub Łucki, Russell Campbell, Asankhaya Sharma, Khalida Meer, Shreen Gul, Daniel Espinosa Gonzalez, Xavier Alapont, Alex Hoover, Gunjan Chhablani, Freddie Vargus, Arunim Agarwal, Yibo Jiang, Deepakkumar Patil, David Outevsky, Kevin Joseph Scaria, Rajat Maheshwari, Abdelkader Dendane, Priti Shukla, Ashley Cartwright, Sergei Bogdanov, Niels Mündler, Sören Möller, Luca Arnaboldi, Kunvar Thaman, Muhammad Rehan Siddiqi, Prajvi Saxena, Himanshu Gupta, Tony Fruhauff, Glen Sherman, Mátyás Vincze, Siranut Usawasutsakorn, Dylan Ler, Anil Radhakrishnan, Innocent Enyekwe, Sk Md Salauddin, Jiang Muzhen, Aleksandr Maksapetyan, Vivien Rossbach, Chris Harjadi, Mohsen Bahaloohoreh, Claire Sparrow, Jasdeep Sidhu, Sam Ali, Song Bian, John Lai, Eric Singer, Justine Leon Uro, Greg Bateman, Mohamed Sayed, Ahmed Menshawy, Darling Duclosel, Dario Bezzi, Yashaswini Jain, Ashley Aaron, Murat Tiryakioglu, Sheeshram Siddh, Keith Krenek, Imad Ali Shah, Jun Jin, Scott Creighton, Denis Peskoff, Zienab EL-Wasif, Ragavendran P, Michael Richmond, Joseph McGowan, Tejal Patwardhan, Hao-Yu Sun, Ting Sun, Nikola Zubić, Samuele Sala, Stephen Ebert, Jean Kaddour, Manuel Schottdorf, Dianzhuo Wang, Gerol Petruzella, Alex Meiburg, Tilen Medved, Ali ElSheikh, S Ashwin Hebbar, Lorenzo Vaquero, Xianjun Yang, Jason Poulos, Vilém Zouhar, Sergey Bogdanik, Mingfang Zhang, Jorge Sanz-Ros, David Anugraha, Yinwei Dai, Anh N. Nhu, Xue Wang, Ali Anil Demircali, Zhibai Jia, Yuyin Zhou, Juncheng Wu, Mike He, Nitin Chandok, Aarush Sinha, Gaoxiang Luo, Long Le, Mickaël Noyé, Michał Perełkiewicz, Ioannis Pantidis, Tianbo Qi, Soham Sachin Purohit, Letitia Parcalabescu, Thai-Hoa Nguyen, Genta Indra Winata, Edoardo M. Ponti, Hanchen Li, Kaustubh Dhole, Jongee Park, Dario Abbondanza, Yuanli Wang, Anupam Nayak, Diogo M. Caetano, Antonio A. W. L. Wong, Maria del Rio-Chanona, Dániel Kondor, Pieter Francois, Ed Chalstrey, Jakob Zsambok, Dan Hoyer, Jenny Reddish, Jakob Hauser, Francisco-Javier Rodrigo-Ginés, Suchandra Datta, Maxwell Shepherd, Thom Kamphuis, Qizheng Zhang, Hyunjun Kim, Ruiji Sun, Jianzhu Yao, Franck Dernoncourt, Satyapriya Krishna, Sina Rismanchian, Bonan Pu, Francesco Pinto, Yingheng Wang, Kumar Shridhar, Kalon J. Overholt, Glib Briia, Hieu Nguyen, David, Soler Bartomeu, Tony CY Pang, Adam Wecker, Yifan Xiong, Fanfei Li, Lukas S. Huber, Joshua Jaeger, Romano De Maddalena, Xing Han Lù, Yuhui Zhang, Claas Beger, Patrick Tser Jern Kon, Sean Li, Vivek Sanker, Ming Yin, Yihao Liang, Xinlu Zhang, Ankit Agrawal, Li S. Yifei, Zechen Zhang, Mu Cai, Yasin Sonmez, Costin Cozianu, Changhao Li, Alex Slen, Shoubin Yu, Hyun Kyu Park, Gabriele Sarti, Marcin Briański, Alessandro Stolfo, Truong An Nguyen, Mike Zhang, Yotam Perlitz, Jose Hernandez-Orallo, Runjia Li, Amin Shabani, Felix Juefei-Xu, Shikhar Dhingra, Orr Zohar, My Chiffon Nguyen, Alexander Pondaven, Abdurrahim Yilmaz, Xuandong Zhao, Chuanyang Jin, Muyan Jiang, Stefan Todoran, Xinyao Han, Jules Kreuer, Brian Rabern, Anna Plassart, Martino Maggetti, Luther Yap, Robert Geirhos, Jonathon Kean, Dingsu Wang, Sina Mollaei, Chenkai Sun, Yifan Yin, Shiqi Wang, Rui Li, Yaowen Chang, Anjiang Wei, Alice Bizeul, Xiaohan Wang, Alexandre Oliveira Arrais, Kushin Mukherjee, Jorge Chamorro-Padial, Jiachen Liu, Xingyu Qu, Junyi Guan, Adam Bouyamourn, Shuyu Wu, Martyna Plomecka, Junda Chen, Mengze Tang, Jiaqi Deng, Shreyas Subramanian, Haocheng Xi, Haoxuan Chen, Weizhi Zhang, Yinuo Ren, Haoqin Tu, Sejong Kim, Yushun Chen, Sara Vera Marjanović, Junwoo Ha, Grzegorz Luczyna, Jeff J. Ma, Zewen Shen, Dawn Song, Cedegao E. Zhang, Zhun Wang, Gaël Gendron, Yunze Xiao, Leo Smucker, Erica Weng, Kwok Hao Lee, Zhe Ye, Stefano Ermon, Ignacio D. Lopez-Miguel, Theo Knights, Anthony Gitter, Namkyu Park, Boyi Wei, Hongzheng Chen, Kunal Pai, Ahmed Elkhanany, Han Lin, Philipp D. Siedler, Jichao Fang, Ritwik Mishra, Károly Zsolnai-Fehér, Xilin Jiang, Shadab Khan, Jun Yuan, Rishab Kumar Jain, Xi Lin, Mike Peterson, Zhe Wang, Aditya Malusare, Maosen Tang, Isha Gupta, Ivan Fosin, Timothy Kang, Barbara Dworakowska, Kazuki Matsumoto, Guangyao Zheng, Gerben Sewuster, Jorge Pretel Villanueva, Ivan Rannev, Igor Chernyavsky, Jiale Chen, Deepayan Banik, Ben Racz, Wenchao Dong, Jianxin Wang, Laila Bashmal, Duarte V. Gonçalves, Wei Hu, Kaushik Bar, Ondrej Bohdal, Atharv Singh Patlan, Shehzaad Dhuliawala, Caroline Geirhos, Julien Wist, Yuval Kansal, Bingsen Chen, Kutay Tire, Atak Talay Yücel, Brandon Christof, Veerupaksh Singla, Zijian Song, Sanxing Chen, Jiaxin Ge, Kaustubh Ponkshe, Isaac Park, Tianneng Shi, Martin Q. Ma, Joshua Mak, Sherwin Lai, Antoine Moulin, Zhuo Cheng, Zhanda Zhu, Ziyi Zhang, Vaidehi Patil, Ketan Jha, Qiutong Men, Jiaxuan Wu, Tianchi Zhang, Bruno Hebling Vieira, Alham Fikri Aji, Jae-Won Chung, Mohammed Mahfoud, Ha Thi Hoang, Marc Sperzel, Wei Hao, Kristof Meding, Sihan Xu, Vassilis Kostakos, Davide Manini, Yueying Liu, Christopher Toukmaji, Jay Paek, Eunmi Yu, Arif Engin Demircali, Zhiyi Sun, Ivan Dewerpe, Hongsen Qin, Roman Pflugfelder, James Bailey, Johnathan Morris, Ville Heilala, Sybille Rosset, Zishun Yu, Peter E. Chen, Woongyeong Yeo, Eeshaan Jain, Ryan Yang, Sreekar Chigurupati, Julia Chernyavsky, Sai Prajwal Reddy, Subhashini Venugopalan, Hunar Batra, Core Francisco Park, Hieu Tran, Guilherme Maximiano, Genghan Zhang, Yizhuo Liang, Hu Shiyu, Rongwu Xu, Rui Pan, Siddharth Suresh, Ziqi Liu, Samaksh Gulati, Songyang Zhang, Peter Turchin, Christopher W. Bartlett, Christopher R. Scotese, Phuong M. Cao, Ben Wu, Jacek Karwowski, Davide Scaramuzza, Aakaash Nattanmai, Gordon McKellips, Anish Cheraku, Asim Suhail, Ethan Luo, Marvin Deng, Jason Luo, Ashley Zhang, Kavin Jindel, Jay Paek, Kasper Halevy, Allen Baranov, Michael Liu, Advaith Avadhanam, David Zhang, Vincent Cheng, Brad Ma, Evan Fu, Liam Do, Joshua Lass, Hubert Yang, Surya Sunkari, Vishruth Bharath, Violet Ai, James Leung, Rishit Agrawal, Alan Zhou, Kevin Chen, Tejas Kalpathi, Ziqi Xu, Gavin Wang, Tyler Xiao, Erik Maung, Sam Lee, Ryan Yang, Roy Yue, Ben Zhao, Julia Yoon, Sunny Sun, Aryan Singh, Ethan Luo, Clark Peng, Tyler Osbey, Taozhi Wang, Daryl Echeazu, Hubert Yang, Timothy Wu, Spandan Patel, Vidhi Kulkarni, Vijaykaarti Sundarapandiyan, Ashley Zhang, Andrew Le, Zafir Nasim, Srikar Yalam, Ritesh Kasamsetty, Soham Samal, Hubert Yang, David Sun, Nihar Shah, Abhijeet Saha, Alex Zhang, Leon Nguyen, Laasya Nagumalli, Kaixin Wang, Alan Zhou, Aidan Wu, Jason Luo, Anwith Telluri, Steven Dillmann, Zhengxiang Wang, Junyu Luo, Hugo Lunn, Artem Gazizov, Haitz Sáez de Ocáriz Borde, Ivan Trus, Morgan Hervault, Zheyu Zhang, Bo Chen, Yuchen Wu, Christopher J. Cordier, Gün Kaynar, Cansin Ayvaz, Polina Avdiunina, Johannes Brust, Xingjian Diao, K. D. Meaney, Yifan Gu, Chenyu Wang, Chenzhuo Dong, William Wright, Simon Brave, Owen Root, Jiayuan Liu, Chow Chun Lok, Tianqin Li, Shiyi Du, Dailan He, Lufeiya Liu, Sina Jamalzadegan, Anil Ramakrishna, Xuanqing Xu, Xin Qing, Xin Luo, Wenkai Li, Shi Bo, Filipp Gusev, Maximos Skandalis, Desheng Ma, Chunhui Zhang, Haoran Qiu, Allen G Hart, Rickard Brüel Gabrielsson, Ido Akov, Artem Lukoianov, Summer Yue, Alexandr Wang, Dan Hendrycks

arXiv:2501.14249v11cs.LGcs.AIcs.CL

TL;DR

Existing benchmarks are saturating, limiting precise measurement of frontier LLM capabilities. The paper introduces HLE, a globally expert-developed, multi-modal benchmark of 2,500 difficult questions, and finds that frontier models remain inaccurate and poorly calibrated on it.

  • Problem

    Popular benchmarks such as MMLU exceed 90% accuracy, limiting precise measurement of rapidly improving frontier LLM capabilities.

  • Method

    HLE combines 2,500 expert-developed multi-modal questions across broad subjects with multiple-choice and exact-match formats, public release, and a private held-out set.

  • Results

    Frontier LLMs consistently achieve low accuracy on HLE and show RMS calibration errors above 70% across models.

  • Takeaways & Limitations

    HLE provides a common reference point for scientists and policymakers to assess AI capabilities, trajectories, risks, and governance measures.

  • Takeaways & Limitations

    High HLE accuracy would indicate expert-level performance on closed-ended, verifiable questions but would not alone establish autonomous research capabilities or artificial general intelligence.

Abstract

from arXiv · show

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 over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind 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 via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant 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.

1 Center for AI Safety, 2 Scale AI

The paper credits named contributors, late contributors, auditors, and a rolling contributor group.

  • Named contributors include Dmitry Dodonov, Tung Nguyen, and Daron Anderson.
  • The paper separately lists late contributors and auditors.
  • HLE-Rolling Contributors are listed in Appendix A.

1 Introduction

Existing benchmarks have become saturated, limiting precise measurement of frontier LLM capabilities. HLE addresses this gap with a difficult, expert-developed benchmark, on which frontier models remain inaccurate and poorly calibrated.

  • Over 90% accuracy on MMLU and similar benchmarks limits precise measurement of frontier LLM capabilities.
  • HLE contains 2,500 multi-modal questions across dozens of subjects, using multiple-choice and exact-match formats resistant to simple retrieval.
  • Questions are tested against state-of-the-art LLMs, iteratively reviewed by graduate-level reviewers, and approved by organizers or experts.
  • Frontier LLMs show low accuracy and RMS calibration errors above 70% across models on HLE.
  • HLE publicly releases 2,500 questions while retaining a private held-out set to assess model overfitting.
  • Figure 1 compares HLE accuracy with the saturation of existing benchmarks across several frontier models.

2 Related Work

Benchmarks measure LLM capabilities across academic, technical, and assistance tasks, but many existing evaluations are approaching saturation. Related work therefore explores harder, multi-modal, expert-authored, and multi-stage-reviewed benchmarks.

  • LLM benchmarks evaluate scientific and mathematical reasoning, code generation, and general-purpose assistance using scalable objective formats.
  • Existing evaluations are nearing perfect scores, motivating harder tests with multi-modal tasks, strengthened datasets, expert authorship, and multi-stage review.

3 Dataset

HLE is a globally constructed dataset of 2,500 challenging questions spanning over a hundred subjects and multiple answer formats. Its pipeline filters for model difficulty, applies expert review, and retains a private held-out set.

  • Dataset scope: HLE contains 2,500 challenging questions across over a hundred subjects, with public questions and private held-out questions.
  • Dataset scope: Questions come from nearly 1,000 subject experts affiliated with over 500 institutions across 50 countries.
  • Question design: Figure 2 presents samples of the diverse and challenging questions submitted to HLE.
  • Question design: 14% of questions require text-and-image comprehension, while 24% are multiple-choice and the remainder are exact-match.
  • Question design: Each submission includes the question, answer specification, detailed solution rationale, subject, and contributor affiliation.
  • Question design: Submissions must be original, precise, unambiguous, solvable, and non-searchable, typically requiring graduate-level expertise.
  • Collection incentives: A $500,000 USD prize pool rewards the top 50 questions with $5,000 each and the next 500 with $500 each.
  • Review pipeline: Over 70,000 attempts yielded approximately 13,000 questions that stumped frontier LLMs and advanced to expert review.

4 Evaluation

The evaluation measures frontier LLM accuracy, calibration, and token use on HLE. Models show low accuracy, frequent overconfidence, and substantially higher token use for reasoning models.

  • Evaluation setup: The evaluation analyzes model performance across question types and domains using quantitative metrics.Responses are standardized into explicit reasoning and final-answer formats, then judged against provided answers.
  • Accuracy: All frontier models achieve low accuracy on HLE, leaving substantial room to narrow the gap with expert-level academic capabilities.The dataset intentionally filters many questions that existing models answer correctly, while retaining some noise-driven correct and incorrect guesses.
  • Quantitative results: Table 1 reports accuracy and RMS calibration error across models, including text-only evaluation for non-multimodal models.The table caption characterizes the pattern as low accuracy and high calibration error across models.
  • Calibration: Models frequently provide incorrect answers with high confidence on HLE, failing to recognize when questions exceed their capabilities.Calibration is assessed using model-stated confidence and RMS calibration error.
  • Token counts: Reasoning models generate substantially more completion tokens than non-reasoning models while improving performance.Completion counts include both reasoning and output tokens.

5 Discussion

The discussion frames HLE as a reference point for assessing future AI progress and informing research, policy, and governance. It also stresses that HLE measures closed-ended academic capabilities rather than autonomous research or general intelligence.

  • Future Model Performance: HLE’s scope is structured academic problems, so high accuracy would indicate expert-level closed-ended performance but not autonomous research capabilities or artificial general intelligence.The discussion distinguishes technical knowledge and reasoning from open-ended research and creative problem-solving.
  • Future Model Performance: The authors suggest models could exceed 50% accuracy on HLE by the end of 2025, given rapid historical benchmark progress.This is presented as plausible rather than certain.
  • Future Model Performance: HLE may be the last academic exam needed for models, but it is not the last benchmark for AI.The claim is limited to academic exams and does not extend to benchmarking generally.
  • Impact: HLE provides a common reference point for scientists and policymakers to assess AI capabilities, development trajectories, risks, and governance measures.The benchmark is intended to support more informed discussions about these issues.

A.1 Data Contributors & Affiliations

The contributor and affiliation listings document broad participation by named contributors, auditors, and researchers associated with numerous universities and research institutions.

  • Contributors: The contributor lists include late contributors as a distinct group.The late-contributor section contains additional named participants and affiliations.
  • Auditors: Auditors are identified separately, with their work conducted while at 2Scale AI.The auditor designation appears alongside the contributor and affiliation listings.
  • Affiliations: Affiliation entries identify participating organizations across North America, Europe, Asia, South America, and Australia.The listed institutions include the University of Toronto, National University of Singapore, University of Copenhagen, University of Buenos Aires, and University of Western Australia.

B Dataset

HLE’s dataset construction combines frontier-model difficulty screening, structured automated evaluation, post-release review, audits, and searchable-question filtering. Late contributions also produced a second held-out private set for future evaluations.

  • Difficulty screening: Questions are screened against frontier LLMs before submission, with exact-match questions required to stump all tested models and multiple-choice questions required to stump all but one.Testing uses multimodal models for text-and-image questions and adds non-multimodal models for text-only questions.
  • Evaluation: A standardized prompt separates reasoning from final answers, and an automated GPT-4O judge evaluates correctness against provided answers.The evaluation targets structured response formatting and answer verification.
  • Late contributions: Late contributions produced thousands of submissions and a second held-out private set for future evaluations.Organizers manually reviewed the submissions and reported similar difficulty and quality to the initial dataset.
  • Review and refinement: Reviewers were not expected to fully verify specialized solution rationales when verification would take more than five minutes.They instead focused on whether questions aligned with the submission guidelines.
  • Review and refinement: A post-release community bug bounty identified major question or label errors, which organizers manually verified and removed when appropriate.Original question authors were consulted when appropriate.
  • Audit: Auditors fully solved a sample of HLE questions, routing flagged errors among organizers, authors, and auditors until consensus was reached.Audit data was used to further refine the dataset.
  • Searchable questions: Potentially searchable questions were manually audited and removed when they were easily found via web search.The procedure used GPT-4o and Perplexity search models, and observed similar frontier-model performance before and after filtering.

B.3 Expert Disagreement Rate and HLE-Rolling

HLE uses multi-reviewer auditing and author rebuttals to assess difficult, expert-authored questions, while HLE-Rolling will incorporate post-release feedback and new questions.

  • Expert Disagreement Rate: Multiple reviewers and author rebuttals are used to resolve disagreements and assess question validity.Reviewers may identify critical background information needed to confirm an answer.
  • Expert Disagreement Rate: HLE includes questions based on contributors’ hands-on research experiences, testing knowledge beyond readily indexed internet sources.
  • Expert Disagreement Rate: Multiple-choice questions are evaluated by the relative plausibility of their provided options rather than as open-ended searches for a perfect solution.
  • HLE-Rolling: HLE-Rolling is a post-release dataset fork that will be updated with community feedback and new questions.Updates will be made publicly available at lastexam.ai.
  • HLE-Rolling: HLE contains over a hundred subjects overall, with the paper presenting the fifty most popular subjects.

C Evaluation

The evaluation uses structured judging to extract model answers, assess correctness against reference answers, and record reasoning and confidence.

  • Evaluation: The requested model response format contains an explanation, chosen answer, and confidence score between 0% and 100%.
  • Evaluation: The evaluation judge compares each model response with the precise reference answer and determines whether the answers match.
  • Evaluation: The structured judge extracts a final answer, reasoning, correctness, and confidence from each model output.
  • Evaluation: Numerical answers are accepted when they fall within a small margin of error; otherwise, inconsistencies or ambiguity are marked incorrect.

C.2 Text-Only Results

Table 2 reports accuracy and RMS calibration error for models evaluated on HLE’s text-only questions.

  • C.2 Text-Only Results: Table 2 reports model accuracy on HLE’s text-only questions.
  • C.2 Text-Only Results: Table 2 reports RMS calibration error for models on HLE’s text-only questions.
  • C.2 Text-Only Results: The table covers models from Table 1.

C.3 Categorical Results

The paper presents category-wise model performance on HLE and separately reports average output token counts for non-reasoning models.

  • C.3 Categorical Results: Table 3 provides a category-wise breakdown of model performance on HLE.
  • C.4 Non-Reasoning Model Token Counts: Figure 6 reports average output token counts for non-reasoning models.
  • C.3 Categorical Results: Table 4 lists the evaluated model versions and their temperature settings.

C.6 Benchmark Difficulty Comparison

HLE evaluates model accuracy with zero-shot chain-of-thought prompts while comparing selected prior-benchmark results from reported sources. Its questions are refined through expert review and screened for difficulty, precision, objective solvability, originality, and broad subject coverage.

  • Evaluation setup: HLE model accuracy is evaluated using zero-shot chain-of-thought prompts, while prior-benchmark results come from reported external sources.The cited examples include zero-shot results for GPT-4O, O1-PREVIEW, and Claude 3.5 Sonnet, and 5-shot MMLU results for Gemini 1.5 Pro.
  • Expert review: Two rounds of human review address whether questions are genuinely difficult rather than merely adversarial to models.Subject-matter reviewers score submissions, provide feedback, and iteratively refine questions; later reviewers also assess first-round feedback.
  • Difficulty and coverage: Questions are generally expected to reach graduate or PhD level, while allowing difficult nonacademic questions across STEM, law, history, psychology, philosophy, trivia, and related fields.The rubric also permits below-graduate questions when models cannot answer them correctly and accepts difficult trivia, game strategy, and cultural questions.
  • Answerability: Accepted questions must have precise, objectively correct, univocal answers that are known or solvable and suitable for reliable evaluation.The rubric rejects subjective prompts, moral questions, and open-ended requests for proofs, explanations, or theories because they cannot be evaluated properly.
  • Question quality controls: Questions should be original rather than derived from textbooks or Google, and non-standard jargon must be explained.The review rubric also checks language, formatting, numerical precision, LaTeX presentation, and conversion of answerable images into text when possible.
Loading 2501.14249v11…