Weights & Biases

CoreWeave

Product status: Status not recorded

No lifecycle review is recorded for this listing. Its inclusion in the directory does not confirm current availability; check the vendor before relying on it.

No status check date recorded

Experiment tracking and model-development tooling: log runs, sweep parameters, evaluate and share results across a team. No lifecycle review is recorded for this listing. Its inclusion in the directory does not confirm current availability; check the vendor before relying on it.

Product preview

Overview

Weights & Biases is the experiment-tracking tool that grew into a model-development platform: log runs, compare them side by side, sweep hyperparameters, visualise datasets and models, and share the results with a team. The newer layers sit on top of that record — serverless training so you can fine-tune without provisioning GPUs, serverless inference, sandboxes, and an open-source RL framework.

The reason to use it is that the history of what you tried is worth more than any single run. That argument scales with team size and experiment volume: at ten runs a week it is overhead, at a thousand it is the only way anyone knows what has already been tried. The commercial tension is now unusual though, and worth stating plainly — the company is part of CoreWeave, a GPU cloud, so the platform's owner also sells the compute its serverless features run on.

Sources

The factual claims in this review were read from the pages below. Each date is when an editor last opened the source, not when the product changed.

Vendor site Checked 2026-09-19
Vendor documentation Checked 2026-09-19

Pros / Cons

Pros

Turns scattered training runs into a comparable record
Sweeps and dataset visualisation in the same place as the logs
Serverless training removes the GPU-provisioning step for fine-tuning
Strong sharing story for teams, not just individuals

Cons

Overhead is hard to justify for a single researcher with few runs
The platform is now owned by a compute vendor it sells against
The vendor's documentation says its domain changes on 30 September, so migration planning is part of adopting it

Who it's for

Tracking and comparing model training runs

Collaborating on ML experiments across a team

Fine-tuning without managing GPU infrastructure

Evaluating and comparing LLM applications

Bottom line

The default experiment tracker for teams serious about ML, with the caveat that its owner now sells the compute it runs on and its domain is changing.

Key Features

Experiment tracking and comparison

Hyperparameter sweeps

Dataset and model visualisation

Serverless training and inference

Shared team reports

Frequently asked questions

Is Weights & Biases currently available?

No lifecycle review is recorded for this listing. Its inclusion in the directory does not confirm current availability; check the vendor before relying on it.

What is Weights & Biases?

Weights & Biases is a AI Platforms tool developed by CoreWeave. It was introduced in 2017, and its developer is headquartered in San Francisco, USA. Experiment tracking and model-development tooling: log runs, sweep parameters, evaluate and share results across a team.