Agentar-Fin-R1 32B

Parameters

32B

Published

Jul 27, 2025

Agentar-Fin-R1 32B is an AI model developed by Ant Group (China), first published in July 2025. It works in the language domain, on tasks such as financial management, language modeling/generation and question answering.

Epoch AI has no training-compute estimate for this model. The model has 32,000,000,000 parameters. Training ran on 64 NVIDIA A100.

Access: Unreleased. Its weights are not openly released. It is built on top of Qwen3-32B. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Ant Group
Country of organization: China
Domain: Language
Task: Financial management, Language modeling/generation, Question answering
Parameters: 32,000,000,000
Training hardware: NVIDIA A100
Chips used: 64
Training power draw: 50.1 kW
Numerical format: BF16
Model accessibility: Unreleased
Open weights: No
Base model: Qwen3-32B
Epoch confidence: Confident