There's also https://github.com/ggml-org/llama.vim

Which I've been using with Qwen3 Coder. As long as infill is supported, that should work. I'll try later today.