Cangjie-bench - Code Generation (Zero-Shot): leaderboard

Metric: Accuracy (%; share of solutions passing all private unit tests; 155 HumanEval problems written in the Cangjie language; zero-shot, code written from the task alone). Source: arxiv.org. Saturation forecast: Around December 2026. 3 models tracked.

Top models

#ModelScore
1Claude Sonnet 4.545.16
2Qwen 3 Max7.74
3DeepSeek V3.23.23

Interactive version: theaggregate.ai/benchmark?slug=cangjie-bench-code-generation-zero-shot · How It Works · Data refreshed daily, snapshot 2026-09-26.