Agents tested on real computer science research tasks make partial progress
FrontierCS 2: A Benchmark for Computer Science R&D
Machine Learning
Summary
The paper looks at whether AI agents can help with tough computer science research problems. It sets up 48 challenges where the AI must improve existing research code after removing its key ideas. Unlike past tests, these challenges let the AI work within real development projects by submitting code updates over time. The best AI agent completed successful runs on just over half the attempts but struggled with about a third of the tasks. This shows current AI can help but still has limits when doing advanced computer science work.
What this means in practice
- •For software engineering teams: Test AI code agents on complex research codebases to assist with iterative development and enhancement.
- •For machine learning infrastructure teams: Evaluate AI agents' ability to improve machine learning systems within real development environments for system performance gains.
Authors
Qiuyang Mang, Wenhao Chai, Runyuan He, Dongruixuan Li, Zihan Zheng, Yujun He, Xuanyi Zhou, Hao Wang, Siddhant Ray, Bo Peng, Ruyi Ji, Shuo Yang, Haocheng Xi, Boru Chen, Xinyang David Han, Melissa Pan, Mayank Mishra, Boxi Yu, Sijia Liu, Kerem Akillioglu, Yi Pan, Zirui Wang, Mert Cemri, Diogo Antunes, Yichuan Wang, Aiwei Ivy Zhang, Zhenyang Chen, Qizheng Zhang, Kaiyuan Liu, Wenyuan Huang, Sherry Ruan, Shu Liu, Dacheng Li, Zerui Cheng, Zeyu Shen, Shang Zhou, Hanfei Chen, Yuzhong Zhang, Bohan Lyu, Qu Ao, Han Zheng, Chonghe Jiang, Jingzhuo Hu, Xiaokun Chen, Kunjun Li, Zhiqing Zhong, Haoye Tian, Tong Zhao, Yunzhong He, Hanchen Li, Ofir Press, Jiantao Jiao, Kunle Olukotun, Karthik R. Narasimhan, Jingbo Shang, Koushik Sen, Matei Zaharia, Alex Dimakis, Ion Stoica, Joseph E. Gonzalez, Alvin Cheung
Abstract
Can LLM agents make contributions to frontier computer science R&D? We introduce FrontierCS 2, a benchmark of 48 open-ended tasks. We take a paper and the code its authors released and delete the paper's contribution. The deleted code becomes the human reference that sets the bar. Earlier open-ended benchmarks give each agent a problem statement and grade one program after the run. FrontierCS 2 places the agent inside a research environment. The agent inherits a working repository and may submit a patch at any point. For each patch, a judge returns a development-suite score but keeps the evaluator, the held-out final suite and the reference private, and grades the final submission on that suite against the reference. The tasks span 13 domains, from machine-learning systems and high-performance computing to databases, security and robotics. Beyond these 48 tasks, FrontierCS 2 keeps a private set of more than 100 tasks, to be announced in the future. We evaluate 12 agents. The best one is Claude Opus 5.5, passing 58% of its runs. No run passes on 16 tasks. A higher score is not the same as a pass: most runs that beat the reference on the combined final score still fall short on some metric or workload. On a shared capability scale for agents and tasks, the median task is roughly as difficult as what the strongest agent can handle.