AI models
MatterGen
Training compute
2.7×10¹⁹ FLOP
Parameters
46.8M
Published
Jan 16, 2025
MatterGen is an AI model developed by Microsoft Research AI for Science (United States), first published in January 2025. It works in the materials science domain, on tasks such as materials design and crystal discovery.
Training it took an estimated 2.7×10¹⁹ FLOP of compute (estimation method: hardware). The model has 46,800,000 parameters. Training ran on 8 NVIDIA A100 for about 10 hours.
Access: Open weights (unrestricted). Its weights are openly available. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Microsoft Research AI for Science
- Country of organization
- United States
- Domain
- Materials science
- Task
- Materials design, Crystal discovery
- Training compute
- 2.7×10¹⁹ FLOP
- Compute estimation method
- Hardware
- Parameters
- 46,800,000
- Training hardware
- NVIDIA A100
- Chips used
- 8
- Training time
- 10 h
- Training power draw
- 6.3 kW
- Model accessibility
- Open weights (unrestricted)
- Open weights
- Yes
- Epoch confidence
- Confident
More from Microsoft Research AI for Science
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