Street Calls of the Week
Investing.com-- Ant Group on Thursday released a new, open-sourced, trillion-parameter large language artificial intelligence model aimed at competing with comparable offerings from DeepSeek and OpenAI.
The Chinese fintech said the new model– called Ling-1T– was capable of high performance on multiple complex reasoning benchmarks while also having constrained output token limits.
The model delivered “improved results across diverse use cases– including code generation, software development, competition-level mathematics problem solving, and logical reasoning, Ant Group said in a statement.
Chinese media reports said the model outperformed comparable models such as DeepSeek V3.1 Terminus and OpenAI’s GPT-5 on several major coding benchmarks.
The Jack Ma and Alibaba (NYSE:BABA)-backed fintech had first joined the AI race in 2023 with a financial LLM. Ant last month released Ring-1T-preview, another 1 trillion parameter model.
The company ranks among several Chinese technology giants racing to release more models and gain a greater foothold in the rapidly growing AI industry.
Ant was also seen ramping up its investments in China’s semiconductor industry, amid a broader push by Beijing for more locally-sourced AI processors.