It can't run LLMs very well, you'll be limited to tiny models with no coding ability and they'll be slow.
i assumed you're connected to internet and using codex/claude code
i assumed you're connected to internet and using codex/claude code