Qwen 3.6 35B MoE

A hobbyist reports successfully running local AI inference on a Dell E7470 laptop with an OCuLink eGPU adapter and a Radeon Vega 64, achieving 25–40 tokens per second. The setup, which uses 8GB Vega 6…

Benchmarking Qwen 3.6 35B MoE (3B active) on an RTX 3090 using Unsloth's UD-IQ4_NL_XL quantisation achieved up to 140 tokens per second for generation and over 3,300 tok/s for prompt processing with a…