Sold by
Weights & Biases AI Development Platform for AWS
Weights & Biases provides AI developers with the tools needed to build models faster, fine-tune LLMs, and develop GenAI applications with confidence for enterprises of all sizes in any vertical.
Reviews (59)
Muhammed A.
Essential ML Experiment Tracking with Real-Time Metrics and Team Collaboration
Reviewed on Jul 31, 2026
Review provided by G2
What do you like best about the product?
Weights & Biases has become an essential platform for managing machine learning experiments, model training, and performance tracking. The interface makes it easy to compare runs, visualize metrics in real time, and collaborate across teams, while integrations with popular ML frameworks simplify adoption. Experiment tracking, artifact versioning, and reproducibility features significantly reduce manual work, helping teams iterate faster, improve model quality, and maintain organized AI development workflows.
What do you dislike about the product?
Weights & Biases offers a comprehensive feature set, but new users may face a learning curve when configuring advanced experiment tracking, reports, and team workflows. Large projects with thousands of experiment runs can sometimes make dashboards feel cluttered, and premium features may be costly for smaller teams. I would also like to see more customization options for visualizations and reporting, along with additional native integrations for enterprise MLOps environments.
What problems is the product solving and how is that benefiting you?
Before using Weights & Biases, tracking machine learning experiments, comparing model performance, and managing training artifacts across multiple projects was time-consuming and difficult to reproduce. The platform centralized experiment tracking, visualization, model versioning, and collaboration in a single workspace, making it much easier to monitor progress and identify the best-performing models. This has reduced manual effort, improved reproducibility, accelerated model development cycles, and enabled the team to make faster, data-driven decisions throughout the ML lifecycle.
Automotive
Automatic Metrics Tracking, but Overall Experience Needs Improvement
Reviewed on Jul 30, 2026
Review provided by G2
What do you like best about the product?
Automatically records metrics, code versions, making results better
What do you dislike about the product?
Projects can get cluttered over time, and that can feel overwhelming.
What problems is the product solving and how is that benefiting you?
Keeps records and training for every run.
Helps identify changes
Helps identify changes
Biotechnology
ML Experiment Tracking, Forward Deployment, and Open-Weight Models Made Easy
Reviewed on Jul 28, 2026
Review provided by G2
What do you like best about the product?
Makes tracking training experiments and sharing training data with my team easy, with dashboards similar to Tensorboard and low performance overhead. Easy to get started with. Backs up data to the cloud and works from a remote cluster seamlessly. Plus offers support for purchasing cloud compute for LLM fine-tuning and FAAS.
What do you dislike about the product?
It doesn't display large quantities of data well, and it's difficult to use some of the more complex visualizations. As a place for publishing/using models, HuggingFace has a larger library and simpler API. Cloud compute pricing is competitive but higher than competitors.
What problems is the product solving and how is that benefiting you?
It helps us log ML training/evaluation data (though the Experiments and Reports features) remotely as I work on an HPC cluster. I can access the data anytime through the mobile app or website, which is convenient because we don't need a secure connection to the cluster. We can also save model weights/architectures and publish them online alongside our academic papers.
Dhruv P.
Solid MLOps platform for experiment tracking with great collaboration features
Reviewed on Jul 28, 2026
Review provided by G2
What do you like best about the product?
Excellent experiment tracking and visualization dashboard that makes it easy to compare model runs and parameters. Strong integrations with major ML frameworks and seamless team collaboration features. The API is intuitive and well-documented, making it straightforward to log metrics and artifacts.
What do you dislike about the product?
Pricing scales steeply with team size, which can be a barrier for smaller organizations. The learning curve for advanced features like custom dashboards and reports is moderate, and documentation could be more comprehensive for edge cases. Occasional UI/UX inconsistencies across different features.
What problems is the product solving and how is that benefiting you?
Helps organize and track ML experiments systematically, reducing time spent manually managing experiment logs and parameters. Enables better collaboration across teams by centralizing model run history and results. Improves reproducibility and debugging of models by maintaining complete audit trails. Accelerates model iteration cycles and provides visibility into which hyperparameters yield the best performance.
Jeni J.
A Must-Have Tool for Keeping ML Experiments Organized
Reviewed on Jul 28, 2026
Review provided by G2
What do you like best about the product?
I primarily use Weights & Biases to track and compare machine learning experiments, monitor training metrics in real time, and manage model versions. I really like how it solves the challenge of keeping experiments organized and reproducible, with everything logged automatically. What I like most about Weights & Biases is how effortless it makes experiment tracking and visualization. The interactive dashboards, real-time training metrics, hyperparameter comparison tools, and artifact management are great for understanding model performance, reproducing results, and collaborating with teammates without adding much overhead to the workflow. The initial setup was developer friendly too. the AI finetuning, monitoring was very good.
What do you dislike about the product?
One area that could be improved is the onboarding experience for new users, especially when exploring advanced features like Sweeps, Artifacts, and Reports. While the platform is very powerful, it can feel overwhelming at first, so more guided tutorials, in-app tips, and ready-to-use workflow templates would help users become productive much faster. I'd also like to see more flexible dashboard customization and filtering options for large projects with hundreds of experiment runs. Better cost and resource usage insights, along with faster loading times for very large experiment histories, would make the platform even more efficient for teams managing complex machine learning and LLM workflows.
What problems is the product solving and how is that benefiting you?
Weights & Biases solves organizing and reproducing ML experiments, automates tracking metrics, hyperparameters, and versions, and aids collaboration in AI projects. It helps me monitor training metrics, manage model versions, and track experiments effortlessly.
Muhammad O.
A Reliable Platform for Tracking Machine Learning Experiments
Reviewed on Jul 25, 2026
Review provided by G2
What do you like best about the product?
What I like most is how easy it is to get started and keep all my experiments organized in one place. The dashboard feels clean and intuitive, so it’s straightforward to track runs, compare results, and share progress with teammates. Overall, it helps me manage model development in a more structured way without ever feeling overly complicated.
What do you dislike about the product?
The platform offers a lot of features, so it can feel a bit overwhelming when you’re first getting started. It took me some time to figure out where everything was and how it all fit together, but after I spent a little time exploring, it became much easier to navigate.
What problems is the product solving and how is that benefiting you?
Weights & Biases helps me keep machine learning experiments organized by tracking runs, comparing results, and making it easier to see which changes actually improve a model. It saves time, supports collaboration, and makes it much simpler to reproduce past experiments rather than having to start from scratch.
Clarion I.
AI Tracing and Evaluation Made Easy
Reviewed on Jul 22, 2026
Review provided by G2
What do you like best about the product?
AI tracing feature for most AI models and evaluation.
What do you dislike about the product?
The free plan has limited features hence the need to upgrade to ensure one gets all features for deploying AI models.
What problems is the product solving and how is that benefiting you?
AI inference, tracing and evaluation
Anson D.
Weights & Biases Makes Experiment Tracking and Run Comparisons Effortless
Reviewed on Jul 11, 2026
Review provided by G2
What do you like best about the product?
What I like most about Weights & Biases is how easy it is to keep track of experiments in one place. The dashboard is well organized and makes it simple to compare runs, monitor metrics, and visualize results. It saves a lot of time compared to manually recording experiment details.
What do you dislike about the product?
The platform has a lot of features, so it can feel a bit overwhelming when you're getting started. It took me some time to understand where everything was. Apart from that, I haven't faced any major issues while using it.
What problems is the product solving and how is that benefiting you?
Weights & Biases helps me organize and track machine learning experiments instead of managing everything manually. Having metrics, logs, and experiment history in one dashboard makes it much easier to compare results and understand what changes are improving the model. It has made my workflow more organized and efficient.
Kumar S.
Easy Experiment Tracking and Smooth PyTorch Lightning Integration
Reviewed on Jul 09, 2026
Review provided by G2
What do you like best about the product?
It has easy experiment tracking and smooth integration with tools like PyTorch Lightning, which makes logging metrics and comparing runs very simple
What do you dislike about the product?
The tracked-hours pricing can become expensive when running multiple experiments in parallel. The dashboard also slows down with large logs, and offline sync isn't always reliable after interrupted runs. Improving performance, sync stability, and making pricing more predictable would make the overall experience much better.
What problems is the product solving and how is that benefiting you?
Earlier, we relied on Excel sheets and screenshots to track experiments, which made comparing models and managing runs quite messy. Now W&B automatically logs metrics, hyperparameters, resource usage, and predictions in one dashboard. Comparing runs is much easier, the whole team has better visibility, reports are easy to share, and hyperparameter sweeps have saved us a lot of manual effort while making experiments more reproducible.
Punit Jain
Experiment tracking has transformed model tuning and now supports faster, more informed AI workflows
Reviewed on Jun 30, 2026
Review provided by PeerSpot
What is our primary use case?
I use Weights & Biases primarily for experiment tracking, logging metrics such as loss and accuracy, learning rate, and other parameters. It helps in visualizing training progress in real time, particularly for deeper projects involving dataset modeling, CI/CD pipelines, and similar tasks.
I used Weights & Biases in my personal project involving self-generating adversarial networks, where I tracked generator and discriminator losses over time, logged sample outputs, and compared architectures and hyperparameters. Those losses helped me analyze my model to optimize it so that they became negligible or minimal. Weights & Biases provides real-time dashboards, image logging, experiment comparison, and other useful features.
One thing I appreciate about using Weights & Biases is how it fits into the bigger picture of ML workflows. As a developer, I can integrate end-to-end workflow integrations, which include data pipelines to track, model registry to store and manage, and deployment monitoring, so that I can analyze how models are performing, the losses, and the gains. It also supports hyperparameter tuning and model comparisons, including the comparison of losses and gains. In my opinion, it is a research notebook experiment collaboration tool.
Using Weights & Biases gave me the ability for faster development and also saved my time since analyzing the discriminator and generator losses, which would have taken a lot of time if I did it manually, was done very easily with Weights & Biases. The graphs it provided were also very helpful in analyzing the gains and losses and the accuracy of the generator and discriminator model.
What is most valuable?
The best features that Weights & Biases offers include experiment tracking to monitor accuracy, loss, and learning rate in real time, visualizing the training process with dashboards. Additionally, it automates testing of hyperparameter configurations for different models through the hyperparameter feature. Weights & Biases has a model registry to version data, models, prompts, and code, ensuring reproducibility by linking experiments. Weights & Biases also has an integration ecosystem that works seamlessly with frameworks like PyTorch and TensorFlow.
What needs improvement?
Deployment and monitoring stands out as a feature I wish had further improvement. When I used it, it served as a fine-tuned model directly from Weights & Biases, providing automations for CI/CD pipelines and machine learning.
From my perspective, I don't think Weights & Biases needs significant improvement, but areas involving more image tracking and additional integrations with tools like PyTorch or TensorFlow would be beneficial. I would prefer some AI tools to be integrated, such as Vercel or Netlify for deployments, as that would create ease of use for developers.
In terms of Weights & Biases's AI capabilities, I believe improvements can be made regarding governance and security. In the AI world, many organizations struggle with securing their codes effectively, so if Weights & Biases introduced features related to security score levels, it would be helpful in enhancing security and strengthening code in a cohesive manner.
For how long have I used the solution?
I have been using Weights & Biases since last year.
What do I think about the stability of the solution?
Weights & Biases is stable.
Which solution did I use previously and why did I switch?
Before using Weights & Biases, I had only developed in personal projects involving self-generating adversarial networks, where I analyzed the performance of generator and discriminator models along with their gains and losses.
What was our ROI?
I have seen a return on investment in terms of time saved, with improved accuracy, reduced losses, and increased gains.
What's my experience with pricing, setup cost, and licensing?
I only use the free tier of Weights & Biases, and I do not have any information regarding the prices or setup costs for licensing. The free tier is sufficient for developers who are in college, in my opinion.
Which other solutions did I evaluate?
I only use Weights & Biases and did not evaluate any other options.
What other advice do I have?
I suggest avoiding making the interview too lengthy, as it is meant for review purposes and should not take up thirty minutes to one hour. My overall review rating for Weights & Biases is eight out of ten.