Benchmark RadarAI BENCHMARK PROFILE
OpenHarmony Bench
OpenHarmony Bench evaluates coding agents on 153 app-level ArkTS tasks across three input sources, with 242 feature points and device-based verification.
- Released
- 2026-08-17
- Readiness
- Inspectable
- Primary field
- General AI
Why it matters
It measures end-to-end app-level correctness, filling the gap between function-level coding benchmarks and real-world app development.
Motivation
We present OPENHARMONY BENCH, an app-level coding benchmark for evaluating LLM-based coding agents on OpenHarmony ArkTS applications.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.