In August 2025, Beijing Deep Logical Intelligence Technology Co., Ltd. launched LLaSO, the world's first fully open-source research framework for speech language models (LSLM), providing data, benchmarks, and models to drive open-source collaboration and innovation in the LSLM field.