Build with AI·Standings
Best Model Registries in 2026
Model registries help teams keep track of machine-learning models and the files around them. This list compares free plans, starting prices, model versioning, approval workflows, model lineage, deployment tracking, model aliases and artifact storage.
Check where each service runs and how you can access it. Across the 38 entries, 36 offer a web option, 27 offer an API, 23 can be self-hosted, 15 run on Linux, and nine each support macOS or Windows. Look for a free plan or trial if you want to try a service before paying.
- 38 platforms
- 3 with a public API on this page
- 16 with a free tier
- Checked 16 Oct 2026
- 1 C
Cometcomet.comExperiment platform with API, web, desktop and self-hosted options; free plan and paid tiers.
- 2 DNDatabricks Notebooksdatabricks.com
Notebook platform with web, desktop, mobile and API access; free plan and 14-day trial.
- 3 WB
Weights & Biaseswandb.aiAI platform with experiment tracking, visualizations and hyperparameter sweeps; free plan available.
- 4 D
DagsHubdagshub.comPlatform for datasets, annotations, experiments, models and pipelines; hosts data but not compute.
- 5 C
ClearMLclear.mlAI platform with tools for compute, model work, pipelines, repositories, and LLM deployment.
- 6 GP
GitLab Package Registrydocs.gitlab.comDependency store for private or public packages that other projects can use; free plan available.
- 7 AS
Amazon SageMaker Autopilotaws.amazon.comModel-building service that creates, trains, and tunes models from supplied data in SageMaker Canvas.
- 8 TATrueFoundry AI Gatewaytruefoundry.com
AI gateway offering one API for models across 35+ platforms with routing, governance and controls.
- 9 Z
ZenMLzenml.ioPipeline orchestrator that runs Python AI workflows across cloud services, Kubernetes or laptops.
- 10 HA
Hopsworks AI Lakehousehopsworks.aiHopsworks describes its AI Lakehouse as a platform for building batch, real-time, and LLM AI systems on open standards.
- 11 M
MLRunmlrun.orgOpen-source framework automating data, model and LLM lifecycles across hybrid and multicloud setups.
- 12 JHJozu Hubjozu.com
Registry for models, agents, MCPs and skills with scanning, governance and hosted or self-hosted use.
- 13 MMMLflow Model Registrymlflow.org
Free model store, API and UI for versions, lineage, tags, aliases and deployment workflows.
- 14 P
Polyaxonpolyaxon.comKubernetes control plane for ML, inference and agent workloads with tracking, orchestration and monitoring.
- 15 D
DataRobotdatarobot.comAI platform for developing, deploying and governing predictive, generative and agentic systems.
- 16 SCSnowflake CoCosnowflake.com
CoCo is an AI coding agent for data engineering, analytics, machine learning, and agent-building work in Snowflake.
- 17 KN
Kubeflow Notebookskubeflow.orgOpen-source notebook environments for AI and data workloads running on Kubernetes.
- 18 JMJFrog MLjfrog.com
Machine learning platform for building, deploying, monitoring, and automating model pipelines.
- 19 OD
Oracle Databaseoracle.comSelf-hosted Oracle database with a free AI Database 26ai plan and no stated storage, use, or time limits.
- 20 V
Valohaivalohai.comMLOps platform for orchestrating workflows and managing compute, data, pipelines, and experiments.
- 21 SVSAS Visual Statisticssas.com
Statistics software with analysis tests, regression models, and scripting support.
- 22 OCOracle Code Assistoracle.com
AI coding tool with completion, chat, and an agent for JetBrains IntelliJ IDEA and Visual Studio Code.
- 23 RH
Red Hat OpenShift Pipelinesredhat.comCloud-native CI/CD pipelines that automate application delivery on Red Hat OpenShift.
- 24 RARadiant AI Cloud Model Registrydocs.ori.co
Model repository for versioning custom AI models and preparing them for deployment.
- 25 SF
Snowflake Feature Storedocs.snowflake.comSnowflake feature store for batch and streaming features, Python or SQL transformations and point-in-time data.
Places #1, #2, #3, #4, #5 on this board are open to sponsors. Take #1 · Other ways to advertise
How to choose
Match the way your team works
Web access is available for 36 of the 38 registries. An API is listed for 27, while 23 support self-hosting. Linux support appears on 15 entries; macOS and Windows support appear on nine each.
Choose the access style that fits your existing setup, especially if you need a private installation or programmatic access.
Check free plans and trial periods
Fifteen entries have a free plan, and 16 list a free trial. Examples include Comet, GitLab Package Registry and Weights & Biases, each with a free plan; GitLab and Weights & Biases list 30-day trials.
Paid starting points differ. Listed examples range from ClearML at $15 per user per month and Comet at $19 per user per month to ZenML at $999 per month.
Review model controls
The comparison covers model versioning, approval workflows, model lineage, deployment tracking and model aliases. These fields help you see how a registry handles changes, review steps and deployment records.
Check which of these controls your process needs before choosing a plan.
Plan for model files
Artifact storage is one of the comparison points. If your models include files or other packaged outputs, confirm that the registry's storage approach fits your workflow and deployment tracking needs.
The list includes services such as MLflow Model Registry, Jozu Hub and Hopsworks AI Lakehouse, with pricing details shown where provided.
Questions
How many model registries are listed?
The list contains 38 model registries.
How many have a free plan?
Fifteen of the 38 entries have a free plan.
Which platforms are covered?
Across the list, 36 offer web access, 27 an API, 23 self-hosting, 15 Linux support, and nine each for macOS and Windows.
Are any entries open source?
No. The list records zero open-source entries.
What pricing examples are shown?
Examples include ClearML from $15/user/month, Comet from $19/user/month, and ZenML from $999/month. Some entries are free or list a trial.
Some top positions on this board can be paid placements, standing where they were sold. The facts on each platform are not for sale. How we rank · Disclosure