Gemma-SEA-LION-v4-27B-IT

Training compute

2.3×10²⁴ FLOP

Parameters

27B

Published

Aug 20, 2025

Gemma-SEA-LION-v4-27B-IT is an AI model developed by AI Singapore (Singapore), first published in August 2025. It works in the multimodal, language and vision domain, on tasks such as language modeling/generation, question answering and visual question answering.

Training it took an estimated 2.3×10²⁴ FLOP of compute (estimation method: operation counting). The model has 27,000,000,000 parameters. It was trained on roughly 500B datapoints. Training ran on NVIDIA H200 SXM.

Access: Open weights (restricted use). Its weights are openly available. It is built on top of Gemma 3 27B. Epoch AI rates the confidence of this record as confident.

Full record

Organization: AI Singapore
Country of organization: Singapore
Domain: Multimodal, Language, Vision
Task: Language modeling/generation, Question answering, Visual question answering
Training compute: 2.3×10²⁴ FLOP
Compute estimation method: Operation counting
Parameters: 27,000,000,000
Dataset size: 500B
Training hardware: NVIDIA H200 SXM
Numerical format: BF16
Model accessibility: Open weights (restricted use)
Open weights: Yes
Base model: Gemma 3 27B
Epoch confidence: Confident