There is an open issue about adding support for Qwn3 which I have been monitoring, would love to use Qwen3 if possible. Issue - https://github.com/ggml-org/llama.vscode/issues/55
On 128GB I would definitely run a larger model, probably with ~10B active parameters. All depends how many tokens per second is comfortable for you.
To get an idea of the speed difference, there is a benchmark page for llama.cpp on Apple silicon here: https://github.com/ggml-org/llama.cpp/discussions/4167
About quant selection: https://gist.github.com/Artefact2/b5f810600771265fc1e3944228...
And my workaround for 'shortening' prompt processing time: I load the files I want to work on (usually 1-3) into context with the instruction: read the code and wait. And while the LLM is doing the prompt processing I write my instructions of what I want to have done. Usually the LLM is long finished with PP before I am finished with writing instructions. Due to KV caching the LLM then gives almost instantly the answer.
For VSCode I use continue.dev as it allows to set my own (short) system prompt. I get around 50token/sec generation and prompt processing 550t/s.
When giving well defined small tasks, it is as good as any frontier model.
I like the speed and low latency and the availability while on the plane/train or off-grid.
Also decent FIM with the llama.cpp VSCode plugin.
If I need more intelligence my personal favourites are Claude and Deepseek via API.