ResearchBenchmarksLatency Profiling on Mamba2-8B Nano 8G (inference)Follow55.3TPOT (Time Per Output Token)UniQL54.24861.34968.4575.551Dec 3, 2025Evaluation ResultsMethodMethodLinksTPOT (Time Per Output Token)TTLT (Total Latency)UniQLW-bit=4, Prun. p%=35%W-bit=4, Prun. p%=35%2025.1255.328,508.1UniQLW-bit=4, Prun. p%=0%W-bit=4, Prun. p%=0%2025.1281.641,116.3