AI Paper Atlas

Research map / Large language models / Agents, code and math

CUARewardBench: A Benchmark for Evaluating Reward Models for Computer-Using Agents

Haojia Lin, Xiaoyu Tan, Yulei Qin, Zihan Xu, Yuchen Shi, Zongyi Li, Gang Li, Shaofei Cai, Siqi Cai, Yuzheng Cai, Chaoyou Fu, Ke Li and 1 more

ICML 2026 · Large language models › Agents, code and math · approach: Software engineering agents

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use strict unanimous voting instead of majority voting (ablation)Q32-s + G106-s · CUARewardBench ORMNPV: 68.5 -> 84.2Improved
use strict unanimous voting instead of majority voting (ablation)Q32-s + G106-s · CUARewardBench ORMspecificity: 93.2 -> 72.2Hurt
add GLM-4.5V zerogui prompt member to the ensemble (ablation)Q32-s + G106-s (majority) · CUARewardBench ORMNPV: 68.5 -> 84.8Improved
add GLM-4.5V zerogui prompt member to the ensemble (ablation)Q32-s + G106-s (strict unanimous) · CUARewardBench ORMNPV: 84.2 -> 93.3Improved
add GLM-4.5V zerogui prompt member to the ensemble (ablation)Q32-s + G106-s (strict unanimous) · CUARewardBench ORMprecision: 90.1 -> 89.8No clear effect
add GLM-4.5V opencua_reflector prompt member to the ensemble (ablation)Q32-s + G106-s (strict unanimous) · CUARewardBench PRMprecision: 75.6 -> 81.7Improved
add GLM-4.5V opencua_reflector prompt member to the ensemble (ablation)Q32-s + G106-s (strict unanimous) · CUARewardBench PRMNPV: 69.1 -> 85.1Improved
use sewsm prompt instead of zerogui prompt (ablation)Qwen2.5VL-32B · CUARewardBench ORMprecision: 70.9 -> 80.0Improved
use sewsm prompt instead of zerogui prompt (ablation)Qwen2.5VL-32B · CUARewardBench ORMNPV: 76.1 -> 71.7Hurt
use sewsm prompt instead of opencua_reflector prompt (ablation)Qwen2.5VL-32B · CUARewardBench PRMprecision: 60.3 -> 65.3Improved
use sewsm prompt instead of opencua_reflector prompt (ablation)Qwen2.5VL-32B · CUARewardBench PRMNPV: 64.8 -> 62.4Hurt
CUA post-training (GUI-OWL-7B vs Qwen2.5VL-7B)Qwen2.5VL-7B · CUARewardBench ORM (zerogui)overall accuracy: 62.1 -> 68.4Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy