Qwen3 30B (3B active) Thinking-2507

User benchmarks Qwen3-30B-A3B-Thinking-2507-Q4_0 on an AMD MI50 using a specialized llama.cpp-gfx906 fork with flash attention and Q8_0 KV cache quantization. Prompt processing reaches 1224.07 t/s at 512 tokens and generation 63.00 t/s at 128 tokens. Without KV quantization, generation is 79.92 t/s but prompt processing drops to 1167.28 t/s at 512 tokens. The fork is optimized for D=128 head dimension models and reports 5-11% prompt processing improvement from flash attention.