Weights & Biases is an MLOps platform for experiment tracking, model evaluation, dataset versioning, and hyperparameter tuning used by AI teams including OpenAI, Meta, and top foundation model builders.
Pricing
$0-60+ / month
Reviews
1,000+
Founded
2017
Team Size
201-500 employees
Current Deal
Free plan available
About Weights & Biases
Weights & Biases (W&B) provides a comprehensive MLOps platform that enables ML teams to track experiments, version datasets, evaluate model performance, and collaborate on AI projects. The platform is used by 30 foundation model builders including OpenAI, Meta, and Cohere.
The experiment tracking system logs metrics, hyperparameters, code, and artifacts for every training run, enabling reproducibility and comparison across experiments. W&B Sweeps automate hyperparameter optimization, while Reports provide collaborative analysis and visualization tools.
Acquired by CoreWeave in March 2025, W&B continues to serve over 100,000 customers ranging from individual researchers to Fortune 500 AI teams. The platform's focus on the ML experiment lifecycle makes it a critical tool for anyone building and improving AI models.
Pricing
$0-60+ / month
Free: Personal use, unlimited experiments, 100GB storage
Team: Starting at $60/month - Team collaboration, admin controls
Enterprise: Custom pricing - SSO, advanced security, dedicated support
Case Studies
Notable Customers
- OpenAI
- Meta
- Toyota
