Intelligence,
for a better humanity.

At Ant Ling, we build frontier models and make advanced intelligence publicly accessible. We believe AGI should move toward openness, sharing, and scalability—grounded in small, iterative steps that support steady evolution and real-world deployment.

Latest Research

Model Family

Ling

Foundation

General-purpose LLM product line built on an MoE architecture, designed for trillion-parameter scale with low activation cost. It delivers flagship-level intelligence with significantly improved efficiency, spanning models from lightweight to trillion-parameter flagship versions.

Ling-2.5-1T Ling-flash-2.0 Ling-mini-2.0
Model Details
Ring

Reasoning

Large language model product line designed for deep reasoning and cognitive tasks, built on the same trillion-scale MoE architecture as Ling. It delivers low activation cost at massive scale, achieving leading performance among open-source models in mathematics, coding, and complex reasoning tasks.

Ring-2.5-1T Ring-flash-2.0 Ring-mini-2.0
Model Details
Ming

Multimodal

Multimodal model product line designed for any-to-any understanding and generation, built on a unified multimodal architecture. It enables seamless cross-modal intelligence across text, image, audio, and video, delivering strong performance in visual knowledge, immersive speech, and image creation.

Ming-flash-omni-2.0 Ming-UniVision-16B-A3B Ming-UniAudio-16B-A3B-Edit
Model Details
curl --request POST \
  --url https://api.tbox.cn/api/llm/v1/chat/completions \
  --header 'Authorization: Bearer <YOUR_API_KEY>' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "Ling-1T",
    "messages": [
      {
        "role": "user",
        "content": "Hello, Ant Ling"
      }
    ]
  }'

How to Access Ling API

A unified API and Python SDK make integration simple and efficient. From lightweight chatbots to complex autonomous agents, our platform helps your applications go live quickly and scale reliably.