Pleias-RAG-1B

Training compute

3×10²² FLOP

Parameters

1.2B

Published

Apr 25, 2025

Pleias-RAG-1B is an AI model developed by PleIAs (France), first published in April 2025. It works in the language domain, on tasks such as retrieval-augmented generation, language modeling/generation, question answering and 3 more.

Training it took an estimated 3×10²² FLOP of compute (estimation method: operation counting). The model has 1,200,000,000 parameters. Training ran on 16 NVIDIA H100 SXM5 80GB.

Access: Open weights (unrestricted). Its weights are openly available. It is built on top of Pleias 1.0 1.2B. Epoch AI rates the confidence of this record as confident.

Full record

Organization: PleIAs
Country of organization: France
Domain: Language
Task: Retrieval-augmented generation, Language modeling/generation, Question answering, Search, Text summarization, Translation
Training compute: 3×10²² FLOP
Compute estimation method: Operation counting
Parameters: 1,200,000,000
Training hardware: NVIDIA H100 SXM5 80GB
Chips used: 16
Training power draw: 22.0 kW
Model accessibility: Open weights (unrestricted)
Open weights: Yes
Base model: Pleias 1.0 1.2B
Epoch confidence: Confident