The Step-by-Step Stellar Voice Model Reaches the Top of the Global Charts

The large-scale evaluation ranking of Artificial Analysis Speech Reasoning has been updated, and the native voice reasoning model Step-Audio-R1.1 has taken the top spot. This ranking is currently one of the third-party benchmarks for evaluating "native speech models" in the industry. The core consideration is the ability of the model to directly process audio and perform complex logical reasoning, with main dimensions including accuracy, latency, etc.