Industry Adoption

Several major AI companies and research institutions are exploring or already using synthetic data:

1. Anthropic: Used synthetic data in training Claude 3.5 Sonnet.
2. Meta: Fine-tuned Llama 3.1 models with AI-generated data.
3. OpenAI: Reportedly using synthetic data from its "o1" model for the upcoming Orion.
4. Writer: claims to have trained Palmyra X 004 almost entirely on synthetic data at a fraction of the cost of comparable models.
5. Microsoft: Utilized synthetic data in training its Phi open models.