OpenHarmony Bench: Evaluating LLMs and Coding Agents on OpenHarmony App Development
2026-08-17 • Software Engineering
Software Engineering
AI summaryⓘ
The authors created OPENHARMONY BENCH, a test to check how well AI coding helpers can improve whole mobile apps made with OpenHarmony ArkTS, not just small functions. Their test asks the AI to add features, follow detailed instructions, or fix bugs in apps that actually run on devices to see if the changes work. They measured success by whether the app builds without errors and if the requested features behave correctly, finding that while most apps build fine, less than 60% of tasks are fully completed. They also found tasks driven by detailed specs are the hardest for AI to complete. The whole benchmark and its resources are publicly available online.
OpenHarmonyArkTSlarge language modelscoding agentsapp-level benchmarkbuild success ratetask completionfeature requestsbug fixingspec-driven tasks
Authors
Li Li, Han Hu, Tianjian Zhang, Xin Peng, Fangzhu Mao, Qingyu Zhang, Xiaoheng Xie, Zhongmin Tang, Zhihao Lin, Haolin Ruan, Miaomiao Dong, Liuchuan Zhu, Yue Li, Chi Chen, Wenkang Zhong, Mingfei Zhang, Yang Yu, Bo Sun, Chaorui Zhang, Weixi Zhang, Wei Han, Bo Bai, Kui Liu, Gang Fan, Siru Liu, Jiaqian Zhou, Jiali Sun, Yunbiao Dong, Wenhao Zhong, Yunhong Xu
Abstract
We present OPENHARMONY BENCH, an app-level coding benchmark for evaluating LLM-based coding agents on OpenHarmony ArkTS applications. Unlike function-level benchmarks, it evaluates complete app-level changes: each task requires an agent to modify a buildable ArkTS project so that a requested behavior works end to end, involving UI state, data persistence, build configuration, and platform APIs. The benchmark installs and drives the delivered application on a device to check whether the behavior is observable. It covers three input sources: natural-language feature requests (new-feature), structured scenario specifications (spec-driven), and bug descriptions (bug-fix). The benchmark contains 153 top-level tasks and 242 Feature points (F-points), where an F-point is one executable behavior check. The snapshot includes 32 new-feature tasks, 50 spec-driven tasks with 139 F-points, and 71 bug-fix tasks. The main leaderboard is scored over top-level tasks rather than independently weighted F-points. We describe the benchmark construction, statistics, and build-and-test evaluation pipeline, and evaluate DevEco Code with eight LLMs across three independent full-suite runs per configuration. Three findings emerge. First, newer generations complete more tasks than their predecessors within evaluated model-family pairs. Second, buildability is close to saturated while behavioral correctness is not: mean Final Build Success Rate is 94.77% to 100.00%, whereas mean Task Completion is 48.36% to 58.39%. Third, spec-driven tasks have the lowest Task Completion under all-checks task scoring, with no configuration exceeding 35%. The code, data, tasks, reference solutions, tests, evaluation scripts, and leaderboard are released through the official OPENHARMONY BENCH website at https://bench.matrix.openharmony.cn/.