Qwen Next 3.8 / Claude Opus level local model - What to buy in order to deploy?

A Reddit user is seeking advice on hardware for deploying a local AI model, specifically targeting a Qwen 3.8 Next model at Q3 level. They propose a dual GPU, 64GB setup, believing it might be sufficient given that streaming n-grams off RAM/SSD would reduce the total memory requirement from 85GB to 63GB for the model and KV cache. The post emphasizes it is not AI-generated and requests user experience.