Inference
OpenHarmony Bench authors

OpenHarmony Bench v1.0

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

ModelsModels with a result.12
Top resultBest result on this test.60.8%GLM-5.3
Top-3 spreadPoints from first to third.1.8 pts
YearYear of release.2026

Result and price

020406080100$0.1$0.3$1$3$10$30RESULTOUTPUT PRICE PER 1M TOKENS · LOG SCALE

Results

12 results
#ModelResultThe score from the source.
01GLM-5.360.8%
01Qwen3.8 Max60.8%
03DeepSeek V4 Pro 081359.0%
04GLM-5.258.4%
05Kimi K357.3%
05GLM-5.3-Flash57.3%
07Qwen3.8 Max Preview56.0%
08DeepSeek V4 Flash 073153.8%
09Qwen3.7 Max53.4%
10GLM-5.152.3%
11Kimi K2.7 Code52.1%
12MiniMax M348.4%