News

    Ant Group Announces Ling-1T

    Chinese giant Ant Group introduced new Ant Group Ling-1T model, featuring 1 trillion parameters. The model is 10 times faster than Nvidia.

    China-based fintech giant Ant Group has taken a significant step forward in boosting its AI landscape. The company announced the Ant Group Ling-1T, a massive language model with 1 trillion parameters. This move establishes Ant Group as a prominent player in the global artificial general intelligence (AGI) infrastructure market.

    According to the company’s statements, the Ling-1T model demonstrated a high accuracy rate of 70.42% on complex mathematical reasoning tests (2025 AIME). Ant Group states that this model’s performance is comparable to the best existing AI models. This development strengthens the company’s ecosystem, which includes the “Ling” (non-thinking), “Ring” (thinking), and “Ming” (multimodal) series.

    Ant Group Ling-1T

    Ant Group’s announcement wasn’t limited to just the new model. The company also introduced a new inference framework called “dInfer.” This new framework was designed specifically for diffusion language models. However, the most striking claim was that dInfer operates 10 times faster than Nvidia’s solutions.

    The tech giant’s aggressive open-source push is strategically important. Chinese tech companies are trying to overcome the challenges they face in accessing advanced AI chips through algorithmic innovations and software optimizations. Ant Group aims to take this approach and stay ahead of the competition.

    So, what are your thoughts on Ant Group’s new AI initiative? Share your thoughts with us in the comments!

    No comments yet Write the First Comment
    ×

    Your comment has been submitted,
    it will be published after approval.

    Write a Comment