Bering's next-generation native hybrid inference model LING-3.0-flash was officially released. The Ling-3.0 series model implements a systematic upgrade in architecture design, deeply integrating long context computing, state memory, and expert route optimization to carry and transform larger volumes of knowledge and capabilities at a lower single-token activation scale.

Bering's next-generation native hybrid inference model LING-3.0-flash was officially released. The Ling-3.0 series model implements a systematic upgrade in architecture design, deeply integrating long context computing, state memory, and expert route optimization to carry and transform larger volumes of knowledge and capabilities at a lower single-token activation scale.