Yuan 1.0

Training compute

3.5×10²³ FLOP

Parameters

245.7B

Published

Oct 12, 2021

Yuan 1.0 is an AI model developed by Inspur (China), first published in October 2021. It works in the language domain, on tasks such as language modeling. It counts among the frontier models: the systems trained with the most compute of their moment.

Training it took an estimated 3.5×10²³ FLOP of compute (estimation method: reported). The model has 245,730,000,000 parameters. It was trained on roughly 180B datapoints. The compute alone is estimated at $592K in 2023 dollars.

Access: API access. Its weights are not openly released. The reference paper has 69 citations. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Inspur
Country of organization: China
Domain: Language
Task: Language modeling
Training compute: 3.5×10²³ FLOP
Compute estimation method: Reported
Parameters: 245,730,000,000
Dataset size: 180B
Chips used: 2,128
Training cost (2023 USD): $592K
Numerical format: FP16
Model accessibility: API access
Open weights: No
Citations: 69
Epoch confidence: Confident