Performance & Efficiency - Jalapeño demonstrated higher tokens per user compared to the currently available state-of-the-art [1] - Jalapeño achieved greater throughput per kilowatt than the currently available state-of-the-art [1] Industry Benchmarking - Hardware and AI inference performance was evaluated using Semianalysis's InferenceX benchmark [1]