M4 Pro vs M5 Pro (24GB): Qwen 3.8 Flash Next IQ3 with MoEShift at 374 tok/s prefill, 14.7 tok/s decode, 8.5GB peak memory | 消息推送

消息推送
M4 Pro vs M5 Pro (24GB): Qwen 3.8 Flash Next IQ3 with MoEShift at 374 tok/s prefill, 14.7 tok/s decode, 8.5GB peak memory
55分钟前
M4 Pro vs M5 Pro (24GB): Qwen 3.8 Flash Next IQ3 with MoEShift at 374 tok/s prefill, 14.7 tok/s decode, 8.5GB peak memory

🌐 链接: https://www.reddit.com/r/LocalLLaMA/comments/1xrq2fv/m4_pro_vs_m5_pro_24gb_qwen_38_flash_next_iq3_with/
🏷️ 分组: Reddit｜AI·自托管·VPS
🕒 时间: 2026-11-09 20:10:54
Reddit

From the LocalLLaMA community on Reddit

Explore this post and more from the LocalLLaMA community

55分钟前