Tag
A fix for token latency in sequential agent loops is presented, where parallel tool calling improves performance in LLM agent systems.