PART OF YOUR AI ECOSYSTEM
W&B Inference
Weights & Biases Inference — hosted open models alongside the W&B tooling.
ONE CONNECTION. YOUR MODELS.
Use W&B Inference
through BharatRouter.
Connect your own W&B Inference key in Account Settings to route requests on your provider account and negotiated rates.
Your provider key is verified live, encrypted at rest with AES-256-GCM, and never shown again. You can revoke it whenever you need to.
Saving a key also discovers the models your account can serve, including models beyond the public catalog. Use wandb/<model> to address them. Discovered models are visible only to your organization. BYOK requests are free during beta.
EXPLORE THE CATALOG
Models on W&B Inference
Public catalog routes with pricing and model details.
Loading the live catalog…
Catalog rates describe the published routes; BYOK charges follow your provider account. Your discovered wandb/<model> routes are available in your account.
Ready to connect?
Bring your keys. Keep your model options open.