Adding Benchmaxxer Repellant to the Open ASR Leaderboard

The Open ASR Leaderboard has been updated to include the Benchmaxxer Repellant model, which demonstrates significant improvements in automatic speech recognition (ASR) tasks. It achieved a Word Error Rate (WER) reduction of 15% on the Common Voice dataset compared to previous state-of-the-art models. This addition provides practitioners with a new benchmark for evaluating ASR systems and highlights the ongoing advancements in model performance and robustness in real-world applications.