Cangjie-bench - Program Repair (ILA-agent): leaderboard

Metric: Accuracy (%; share of solutions passing all private unit tests; 32 QuixBugs programs ported to Cangjie, each with one bug to fix; ILA-agent: the model explores the official Cangjie documentation and verifies code with the compiler through tools, at most 15 turns). Source: arxiv.org. Saturation forecast: Around December 2026. 3 models tracked.

Top models

#ModelScore
1Claude Sonnet 4.590.63
2Qwen 3 Max81.25
3DeepSeek V3.271.88

Interactive version: theaggregate.ai/benchmark?slug=cangjie-bench-program-repair-ila-agent · How It Works · Data refreshed daily, snapshot 2026-09-26.