I never get the appeal of T3 Chat. I get that you get to use different models in one place but I never had a problem to juggle for cross checking or different use cases. It's advertised to provide a better user experience which I don't see. It claims to be "faster" and I don't see UI rendering is the perf bottleneck in LLM.