Qwen 3.8 Flash-Next on 64GB RAM + 16GB VRAM, worth it over a fully-loaded 3.6 35B-A3B?

A user with a 5080 16GB GPU and 64GB RAM is seeking advice on whether to upgrade from their current Qwen 3.6 35B-A3B setup to Qwen 3.8 Flash-Next. They are currently satisfied with their 3.6 35B-A3B model, which fits entirely within their VRAM using Unsloth quant. The user is asking for experiences from others with similar hardware configurations before committing to a large download.

This post offers a direct comparison of user experience between Qwen 3.6 35B-A3B and Qwen 3.8 Flash-Next on specific hardware, unlike general performance benchmarks.