I have used every AI IDE (Warp, Cursor, Windsurf, Void, etc.) and VS code extension (Cline, Github Copilot, RooCode, Refract, etc.) with every possible model (GPT5, Gemini 2.5 pro, Kimi K2, Qwen, Qwen Coder, Grok 4, Deepseek, Sonic) and nothing compares to Qoder. It nailed the recent API usage something others failed to do due to outdated knowledge base. For example Bevy which updates often (every ~3-5 months) Qoder had no issue correctly using the latest 0.16 API while other models used the outdated 0.15 API. Honestly even on the basic model it’s better than most. Qoder definitely has the edge and potential to be the best AI IDE in my mind it already is. Only problem I’ve had aside from the normal mistakes that all AI models make in introducing bugs is that it needs work executing commands in the terminal and reading the terminal seemed to only happen sometimes and mostly in Quest mode but it would get stuck trying to run a terminal command and not read the output. Happens rarely in Chat mode.