Pass rate vs. lines of code

A quiet scatterplot of Java leaderboard models from Sonar’s LLM code-quality study. Default view compares Google, Anthropic, and OpenAI; use the controls to include all providers or inspect one provider at a time.