Nathan SmithGotta say, I have been quite impressed with MiMo V2 Pro. I have been kicking the tires (via free preview) in OpenCode for a few weeks. So far, it is the only LLM that I have used where the code output is mostly correct.
> "Yeah, this is how I would write it."
For context, I have access to both Claude Code and Cursor. My employer is footing the bill. But using MiMo is the first time AI output actually seems worthwhile, from a "bang for the buck" standpoint.
> "Hmm. I could see myself paying out-of-pocket for this one."
Whereas with both Claude and Cursor, it is more like… Meh. They exist. I can see their merits. But a personal recurring subscription would hardly seem justifiable, given the rework required to produce something usable.
That said, Cursor is absolutely unrivaled in terms of UX in their app. Pressing `Ctrl+L` to send lines of code to the AI agent chat is genius. Beyond that though, the actual quality generated by Cursor (and Claude) tends to be hit-or-miss.
Neil Turner