I built this because I just thought it would be cool to show which LLMs respond faster (between GTP-4o mini, Claude 3 Haiku, and Gemini 2.5 Flash) and show some metrics (TTFT, avg tok/s, total time, and nTokens)