What is Weights & Biases?
Weights & Biases centers on experiment tracking (logging metrics, hyperparameters and artifacts during training), plus a model/artifact registry and, more recently, LLM evaluation and tracing (Weave) for teams building on foundation models. It's used both by individual practitioners on the free tier and by ML teams needing collaboration, artifact lineage and compliance features on paid plans.
Read the full overview
Sources checked 27 September 2026: official website, documentation, pricing.
Weights & BiasesAI Platforms & Generative AIExperiment tracking, model registry and LLM evaluation for ML teams
Key capabilities
- Experiment tracking: Log runs, metrics, hyperparameters and system stats.
- Model & artifact registry: Version datasets and models with lineage.
- Weave: Tracing and evaluation tooling for LLM applications.
- Deployment options: SaaS, self-hosted Server, or a dedicated private cloud on higher tiers.
Free-tier storage and team-seat limits matter once a project moves from a single researcher to a shared team; check current limits before assuming the free plan scales with your team.
“Experiment tracking, model registry and LLM evaluation for ML teams”
Technical details
- Access
- Self-hosted
- Source model
- Commercial
- Founded
- 2017
- Pricing model
- Freemium
- API
- Available
- Website
- wandb.ai ↗
- GitHub
- Source repository ↗
What to verify
Start from the users, systems and operating responsibilities the tool needs to support.
- Confirm current features, licensing and support terms.
- Validate deployment, data location and access control.
- Test integrations, exports and recovery workflows.
Reviews of Weights & Biases
No published reviews yet.
Loading review form…
