Core Insights - Volcano Engine has released a series of updates for the Doubao large model, including Doubao 1.6, which natively supports multiple thinking lengths, and new models such as Doubao Voice Synthesis Model 2.0 and Doubao Voice Replication Model 2.0 [1] Group 1: Model Updates - Doubao 1.6 introduces four thinking lengths (minimum, low, medium, high) to balance model performance, latency, and cost for enterprises, making it the first model in China to support "tiered thinking length adjustment" natively [2] - Doubao 1.6lite is a lighter version of the flagship model, offering faster inference speed and a 53.3% reduction in overall usage costs compared to Doubao 1.5pro in the most commonly used input range of 0-32k [2] - The Smart Model Router, a solution for intelligent model selection, has been launched, allowing automatic selection of the most suitable model for task requests, optimizing both performance and cost [2] Group 2: Market Performance - As of the end of September, the daily token usage for Doubao has exceeded 30 trillion, representing an over 80% increase since the end of May [1] - According to IDC, Volcano Engine holds a 49.2% market share in China's public cloud large model service market, ranking first [1]