Ling-1T

Training compute

6×10²⁴ FLOP

Parameters

1T

Published

Oct 10, 2025

Ling-1T is an AI model developed by Ant Group (China), first published in October 2025. It works in the language domain, on tasks such as language modeling/generation, question answering, quantitative reasoning and 2 more.

Training it took an estimated 6×10²⁴ FLOP of compute (estimation method: operation counting). The model has 1,000,000,000,000 parameters. It was trained on roughly 20T datapoints.

Access: Open weights (unrestricted). Its weights are openly available. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Ant Group
Country of organization: China
Domain: Language
Task: Language modeling/generation, Question answering, Quantitative reasoning, Code generation, Mathematical reasoning
Training compute: 6×10²⁴ FLOP
Compute estimation method: Operation counting
Parameters: 1,000,000,000,000
Dataset size: 20T
Numerical format: FP8
Model accessibility: Open weights (unrestricted)
Open weights: Yes
Epoch confidence: Confident