Ant Group releases the trillion-parameter flagship model Ling-1T and Open Source.

Golden Ten Data reported on October 9 that in the early morning of October 9, Ant Group released the trillion-parameter general language model Ling-1T. According to the official introduction, Ling-1T is the first flagship model of the Ling2.0 series of Ant Bailing, and it is also the largest and most capable non-thinking model launched by the Ant Bailing team so far. The evaluation shows that under the condition of limited output token, Ling-1T achieves SOTA performance in a number of complex reasoning benchmarks. In addition, Ling-1T has achieved leading results in a number of difficult benchmarks such as code generation, software development, competition mathematics, professional mathematics, and logical reasoning, and has ranked first in many indicators of open source models.