AI models
PULI GPTrio
Training compute
5.8×10²¹ FLOP
Parameters
6.7B
Published
Aug 23, 2023
PULI GPTrio is an AI model developed by Hungarian Research Centre for Linguistics (Hungary), first published in August 2023. It works in the language domain, on tasks such as chat and translation.
Training it took an estimated 5.8×10²¹ FLOP of compute (estimation method: hardware). The model has 6,700,000,000 parameters. It was trained on roughly 230.6B datapoints. Training ran on NVIDIA A100 for about 2.2K hours.
Access: Open weights (non-commercial). Its weights are openly available. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Hungarian Research Centre for Linguistics
- Country of organization
- Hungary
- Domain
- Language
- Task
- Chat, Translation
- Training compute
- 5.8×10²¹ FLOP
- Compute estimation method
- Hardware
- Parameters
- 6,700,000,000
- Dataset size
- 230.6B
- Training hardware
- NVIDIA A100
- Training time
- 2,200 h
- Model accessibility
- Open weights (non-commercial)
- Open weights
- Yes
- Epoch confidence
- Confident
SourceEpoch AI, 'AI Models'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/data/ai-models. Licensed under CC BY 4.0.
← All ai models