Who is Comet ML best for?

Comet ML is built for Data scientists and ML engineers building and training models, AI application developers building LLM and GenAI products, MLOps practitioners and DevOps teams managing AI pipelines, Enterprise AI/ML teams requiring compliance and governance. Common use cases include ML experiment tracking and model reproducibility; LLM application debugging, tracing, and root-cause analysis; Agentic workflow observability and evaluation.

How is Comet ML priced?

Opik (LLM observability): Free Open Source (self-hosted, unlimited spans); Free Cloud (up to 10 members, 25k spans/month, 60-day retention); Pro Cloud at $19/month (up to 50 members, 100k spans, additional spans at $5/100k); Enterprise at custom pricing (unlimited members, custom spans, flexible deployment, SSO, SLAs). MLOps Platform: Free (1 user, fair usage, 100GB storage); Pro at $19/user/month (up to 10 users, 1,500 training hours, 500GB storage); Enterprise at custom pricing (unlimited users, unlimited training hours, production monitoring, flexible deployment, SSO). Free Pro plan available for verified academic users across both products.

What are the alternatives to Comet ML?

Common MLOps & Experiment Tracking alternatives to Comet ML include MLflow, Weights & Biases, ClearML, ZenML, Anyscale. See the full comparison hub at /verticals/mlops-experiment-tracking/compare.

What do users praise about Comet ML?

Users frequently praise: Easy integration with major ML frameworks (PyTorch, TensorFlow, Scikit-learn, Keras); Intuitive UI and dashboard visualizations; Minimal setup and boilerplate code required; Responsive Slack customer support; Experiment comparison and reproducibility; Free tier accessible for individuals and academics; Centralized team collaboration on ML experiments.

What are common complaints about Comet ML?

Frequently cited limitations: High enterprise license costs; File upload size limitations causing errors; Inconsistencies between UI features and Python SDK; Difficult to manage large numbers of experiment trails; Learning curve for new users; Documentation could be more polished; Custom view settings can become disorganized.

When was Comet ML founded and where?

Comet ML was founded in 2017, headquartered in New York, USA by Gideon Mendels, Nimrod Lahav.

Comet ML reports 51-200 employees, 150,000+ developers; 10,000+ teams customers.

AI visibility report

Comet ML ranks #4 in MLOps & Experiment Tracking AI search.

Outside the top three on 19 of the 25 prompts buyers actually ask.

MLflow is cited on 17 of those losses.

25 prompts

5 platforms

Updated Sep 17, 2026 - refreshed weekly

Track Comet ML daily

14-day free trial. No card required. Setup comes pre-filled for Comet ML.

Also benchmarked

Comet ML appears in another vertical

AI/ML Infrastructure & LLM Tools

Track Comet ML across these prompts daily.

Start your 14-day trial

5percent

Presence Rate

Low presence

#4 among 6 vendors · still absent from 95.2% of tracked prompt responses

Top-3 citations across 125 prompt × platform pairs

N/A

Sentiment

-1.00.0+1.0

Unknown

#4of 6

Peer Ranking

#1#6

Mid-packin MLOps & Experiment Tracking

Key Metrics

Presence Rate

4.8%

Share of Voice

5.2%

Avg Position

#2.2

Docs Presence

3.2%

Blog Presence

0.0%

Brand Mentions

12.0%

Platform Breakdown

ChatGPT

24%6/25 prompts

Gemini Search

Research dossierCapabilities, use cases, sources, reviews, pricing, and FAQ

Overview

Comet ML (branded as Comet) is a New York-based AI developer platform founded in 2017 by Gideon Mendels and Nimrod Lahav. It offers two flagship product families: an MLOps platform for ML experiment tracking, model registry, dataset management, and production monitoring; and Opik, an open-source LLM observability and evaluation platform launched in September 2024. Trusted by over 150,000 developers and 10,000+ teams—including Netflix, Uber, Etsy, Zappos, and NatWest—Comet supports the full AI development lifecycle. The Opik GitHub repository has surpassed 19,000 stars. The company has raised approximately $70M in total funding, with a $50M Series B led by OpenView in November 2021. Deployments span cloud-hosted, self-hosted, and on-premises options with enterprise-grade compliance.

Comet is an end-to-end AI developer platform with two core product families: (1) a MLOps platform covering experiment tracking, model registry, dataset versioning, and production monitoring for teams building and training ML models; and (2) Opik, a truly open-source LLM observability and evaluation platform for tracing, testing, optimizing, and monitoring LLM applications and agentic workflows—available via self-hosted OSS, managed cloud, or enterprise deployment.

Reviews

4.3/5G2·12+

Praised

Easy integration with major ML frameworks (PyTorch, TensorFlow, Scikit-learn, Keras)
Intuitive UI and dashboard visualizations
Minimal setup and boilerplate code required
Responsive Slack customer support
Experiment comparison and reproducibility
Free tier accessible for individuals and academics
Centralized team collaboration on ML experiments

Criticized

High enterprise license costs
File upload size limitations causing errors
Inconsistencies between UI features and Python SDK
Difficult to manage large numbers of experiment trails
Learning curve for new users
Documentation could be more polished
Custom view settings can become disorganized

G2 reviewers (4.3/5, 12 reviews) and Capterra/Software Advice users praise Comet for easy integration with major ML frameworks, an intuitive dashboard, strong visualization without boilerplate code, and responsive customer support via Slack. Negative themes include high enterprise licensing costs that have led some businesses to consider switching, file upload size limitations causing errors, occasional UI/SDK inconsistencies in the model registry, and difficulty organizing large numbers of experiment trails. Gartner Peer Insights reviewers note that Comet manages the ML lifecycle from training through production with in-depth analytics. Compared to Weights & Biases on G2, reviewers favor W&B on product support quality and roadmap direction.

Prompt	Gemini Search	Bing Copilot	ChatGPT	Google AI Mode	Perplexity
Capability0/5 cited (0%)
I'm evaluating model registries — which platforms offer the best approval workflows, staging environments, and audit trails for enterprise compliance?	Neither your brand nor a competitor was cited	Neither your brand nor a competitor was cited	A competitor was cited	Neither your brand nor a competitor was cited	A competitor was cited
Which MLOps platforms handle multi-modal artifact storage — metrics, model weights, evaluation datasets, and visualizations — without requiring separate tooling?

Developer Experience3/5 cited (60%)
Which ML experiment platforms make it easiest to reproduce a past run exactly, including environment, data version, and hyperparameters?	Neither your brand nor a competitor was cited	Neither your brand nor a competitor was cited	A competitor was cited	Neither your brand nor a competitor was cited	A competitor was cited

Integrations & Ecosystem0/5 cited (0%)
What MLOps platforms have the deepest integrations with object storage backends for versioning large training datasets and model artifacts?	Neither your brand nor a competitor was cited	Neither your brand nor a competitor was cited	A competitor was cited	A competitor was cited	A competitor was cited

Performance & Reliability0/5 cited (0%)
Which self-hosted MLOps platforms have the lowest operational overhead to keep highly available for a 24/7 training pipeline environment?	Neither your brand nor a competitor was cited	Neither your brand nor a competitor was cited	A competitor was cited	A competitor was cited	A competitor was cited

Setup & First Run3/5 cited (60%)
I'm evaluating experiment tracking platforms for a team migrating off a homegrown spreadsheet-based tracking system — what should I look at?	Neither your brand nor a competitor was cited	Neither your brand nor a competitor was cited	Your brand and a competitor were cited	A competitor was cited	A competitor was cited