9h agoBlackbox AI launches its inference engine, running NVIDIA's 550B Nemotron model at 420.2 tokens per secondHosting costs $1.08 per million output tokens.SentimentSentimentPos100%Neg0%Users are excited about NVIDIA's Nemotron Ultra hitting 420+ tokens per second on Blackbox because the impressive inference speed makes it tempting to try out.5 comments with sentiment. View comments.