MarcoPizeta/flash-next-rtxpro6000-bench
Qwen3.8-Flash-Next on a single RTX PRO 6000 Blackwell (96 GB) with 64 GB host RAM: vLLM (NVFP4/FP8 + INT4 PLE offload) vs ExLlamaV3 (EXL3 4.05bpw) - reproducible speed, prefill and tool-calling benchmark
正在读取项目信息…