AI models
HyperCLOVA 82B
Training compute
1.5×10²³ FLOP
Parameters
82B
Published
Sep 10, 2021
HyperCLOVA 82B is an AI model developed by NAVER and Search Solutions (South Korea), first published in September 2021. It works in the language domain, on tasks such as language modeling/generation, chat, translation and text classification.
Training it took an estimated 1.5×10²³ FLOP of compute (estimation method: operation counting,hardware). The model has 82,000,000,000 parameters. It was trained on roughly 300B datapoints. Training ran on 1,024 NVIDIA A100 for about 643 hours. The compute alone is estimated at $586K in 2023 dollars.
Access: API access. Its weights are not openly released. The reference paper has 131 citations. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- NAVER, Search Solutions
- Country of organization
- South Korea
- Domain
- Language
- Task
- Language modeling/generation, Chat, Translation, Text classification
- Training compute
- 1.5×10²³ FLOP
- Compute estimation method
- Operation counting, Hardware
- Parameters
- 82,000,000,000
- Dataset size
- 300B
- Training hardware
- NVIDIA A100
- Chips used
- 1,024
- Training time
- 643 h
- Chip-hours
- 658.6K
- Training power draw
- 827.0 kW
- Training cost (2023 USD)
- $586K
- Model accessibility
- API access
- Open weights
- No
- Citations
- 131
- 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