Weights & Biases is an MLOps platform centered on experiment tracking: logging training runs, metrics, and artifacts so teams can compare configurations and reproduce results. It adds a model registry, hyperparameter sweeps, and the Weave toolkit for LLM applications. A free plan covers personal, non-corporate projects; the Pro plan is a flat monthly subscription with metered storage and ingestion on top, and enterprise plans add dedicated or self-managed deployment. The client SDK is open source while the platform itself is commercial.
experiment-trackingmlopsmodel-registry
Overview
- Founded
- 2017
- Pricing tier
- High
- Startup-friendly
- Yes
- Enterprise-ready
- Yes
ML platform
- Free tier
- Yes: Free plan for personal, non-corporate projects
- Open-source
- No
- Self-hostable
- Yes
- Managed training
- No
- Model hosting
- No
- Fine tuning
- No
- Notebooks
- No