Ant Group has unveiled Ling-1T, a groundbreaking open-source large language model (LLM) boasting an astonishing one trillion parameters. The Chinese fintech giant aims to challenge top-tier AI developers such as OpenAI and DeepSeek with this powerful new model designed for complex reasoning and high-performance computing.
According to Ant Group, Ling-1T excels across a wide range of advanced tasks, including software development, code generation, mathematical problem-solving, and logical reasoning. The company noted that despite its scale, the model maintains efficient output token limits while delivering superior results on multiple reasoning and coding benchmarks.
Chinese media outlets report that Ling-1T surpasses competitors like DeepSeek V3.1 Terminus and OpenAI’s GPT-5 in various coding performance evaluations, marking a significant achievement for China’s rapidly evolving AI ecosystem.
This latest release follows Ant Group’s Ring-1T-preview, another trillion-parameter model introduced just last month. The company, backed by Alibaba (NYSE: BABA) and Jack Ma, first entered the AI race in 2023 with a financial-focused LLM tailored for the fintech industry. Since then, Ant has aggressively expanded its AI research and development to compete on a global scale.
The launch of Ling-1T also aligns with Beijing’s strategic push for technological self-reliance. Ant Group has reportedly been increasing investments in China’s semiconductor industry to support domestic AI infrastructure and reduce dependence on foreign chipmakers.
With Ling-1T, Ant Group reinforces its position among China’s leading tech innovators, signaling a new phase in the global AI race. The model’s open-source release could further accelerate research, collaboration, and the broader adoption of trillion-parameter models within China and beyond.