Is Weights & Biases free? What it actually costs in 2026
Weights & Biases is free for personal work: up to 5 model seats, 5 GB of storage a month and 1 GB of Weave data ingestion. Pro starts at $60 a month for teams, with a 30-day free trial; academics get Pro free.
The Free plan is 'designed for personal development' — experiment tracking, registry and lineage, Weave tracing within 1 GB a month. Storage above the allowance is $0.03 per GB and Weave ingestion above it $0.10 per MB. Pro is for organisations under 50 people; larger ones must move to Enterprise.
What the free tier actually includes
- Up to 5 model seats
- 5 GB of storage a month (then $0.03 per GB)
- 1 GB of Weave data ingestion a month (then $0.10 per MB)
Where the wall is
Weave ingestion. Tracing LLM apps logs every prompt and completion; a gigabyte goes quickly and the overage is $0.10 per megabyte — $100 a GB — which is the number that moves teams to Pro or to a self-hosted tracker.
After that: Pro from $60 a month billed monthly (100 GB storage, 1.5 GB Weave ingestion, unlimited teams); Enterprise custom; additional storage $0.03 per GB, additional ingestion $0.10 per MB; free academic Pro licence with 200 GB and up to 100 seats.
Can you self-host Weights & Biases? Yes
W&B offers a local server under an Enterprise trial licence. MLflow, Aim and ClearML are the open-source experiment trackers compared here, self-hosted on any plan.
What it actually takes to run one yourself
Real RAM against the documented minimum, monthly cost, setup time, and the one thing that catches people out.
Self-hosting MLflow
Self-hosting Aim
Self-hosting ClearML
When paying is still the right call
Stay if the collaboration surface is doing real work — W&B Reports are genuinely good at turning a training run into something a non-ML stakeholder can read, and no open alternative matches that. Stay for large-scale distributed sweeps, where the hosted orchestration is mature and reproducing it is a project. Stay while your team is small enough that per-seat pricing is not the dominant line, and while nobody wants to own a tracking server and its storage. The alternatives below assume you can run a service and a database.
What you are locked into
Moderate. New logging code is a small change — the APIs are similar enough across these tools that swapping is largely mechanical, and MLflow can ingest a good deal of history. The sticky part is accumulated run history and stored artefacts, which is the whole reason the tool exists. Export before you cancel, and be aware that pulling large artefact stores out takes longer than the migration itself.
If you would rather not pay: MLflow
MLflow
The open standard. Tracking, registry, projects and deployment in one.
Free, Apache-2.0. Managed versions are sold by the clouds if you want one.
See all 5 Weights & Biases alternatives comparedCompare the free options head-to-head
Read from Weights & Biases's own pricing and documentation on 2026-09-26 — 25 days ago. Vendors change prices without notice; confirm before you commit.