Tri-21B

Training compute

2.9×10²³ FLOP

Parameters

20.7B

Published

Aug 1, 2025

Tri-21B is an AI model developed by Trillion Labs (South Korea), first published in August 2025. It works in the language domain, on tasks such as language modeling/generation, question answering, code generation and quantitative reasoning.

Training it took an estimated 2.9×10²³ FLOP of compute (estimation method: reported,operation counting). The model has 20,730,000,000 parameters. It was trained on roughly 2.3T datapoints.

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

Full record

Organization: Trillion Labs
Country of organization: South Korea
Domain: Language
Task: Language modeling/generation, Question answering, Code generation, Quantitative reasoning
Training compute: 2.9×10²³ FLOP
Compute estimation method: Reported, Operation counting
Parameters: 20,730,000,000
Dataset size: 2.3T
Model accessibility: Open weights (restricted use)
Open weights: Yes
Epoch confidence: Confident