Weights & Biases
Experiment tracking and model management for machine learning teams.
why this verdict
Keep paying — OpenAI has not replaced it
OpenAI's version overlaps, but does not finish the dev tools job, so this one still earns its $50/mo.
- A named launch, not a vibe named, unlinked
- Fewer teams training models at all as fine-tuning gives way to prompting and retrieval. OpenAI, August 5, 2025 — no announcement link recorded yet.
- How much of the job it covers not the job editorial call
- Parts of it. The job still needs the tool to get finished.
- Is there a free way to do it? yes
- 2 of 3 listed replacements have a usable free tier: Comet and MLflow.
- What the call is worth $600/yr
- $50/mo at the entry paid tier — $600 a year per seat.
- Threatened by
- OpenAI
- Since
- August 5, 2025
- List price
- $50/mo
- Per year
- $600
The backstory
Weights & Biases became the default lab notebook for machine learning, logging runs, metrics, artifacts and hyperparameter sweeps so results are reproducible. The category risk is that far fewer companies train models now than in 2021, since most simply call an API. W&B answered by building Weave for LLM application tracing and evaluation, extending the same discipline to prompting, and its position inside frontier labs that do still train is about as durable as this market gets.
Escape hatches
Comparable experiment tracking with Opik for LLMs
comet.com open_in_newOpen-source tracking you host yourself
mlflow.org open_in_newFocused on very large-scale training runs
neptune.ai open_in_new2 of 3 replacements have a usable free tier.
