Skip to content

OpenHarmony Bench v1.0

An app-level coding benchmark that asks DevEco Code configurations to implement observable behavior in buildable OpenHarmony ArkTS applications.

12 models

Higher is better.

  1. 1Qwen3.8 MaxAlibaba60.8
  2. 2GLM-5.3Z.AI60.8
  3. 3DeepSeek V4.1 Flash 1017DeepSeek60.3
  4. 4DeepSeek V4 Pro 0813DeepSeek59.0
  5. 5GLM-5.2Z.AI58.4
  6. 6Kimi K3Moonshot AI57.3
  7. 7GLM-5.3-FlashZ.AI57.3
  8. 8DeepSeek V4 Flash 0731DeepSeek53.8
  9. 9Qwen3.7 MaxAlibaba53.4
  10. 10GLM-5.1Z.AI52.3
  11. 11Kimi K2.7 CodeMoonshot AI52.1
  12. 12MiniMax M3MiniMax48.4
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Task completion through DevEco Code
Tasks
153 app-development and bug-fix tasks
Authors
OpenHarmony Bench authors (2026)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.