What Comet does

Comet ML, Inc. operates an AI developer platform that combines two product lines on shared infrastructure. The legacy Comet MLOps suite covers the classical ML lifecycle — experiment tracking, dataset and artifact management, model registry, and model production monitoring — through native SDK integrations with PyTorch, TensorFlow, Keras, Hugging Face, scikit-learn, and XGBoost. The newer Opik platform extends the company into LLMOps and AI engineering, providing open-source LLM observability (logging every LLM call, tool invocation, and agent step), automated evaluation against datasets using 30+ LLM-as-a-judge and heuristic metrics, a version-controlled prompt library and playground, and a suite of six prompt/agent optimization algorithms. Modules layered on Opik include Ollie (an autonomous coding agent that turns trace signals into code fixes), Cost Intelligence (visibility and one-click optimization of Claude Code and Codex token spend), and an MCP Server connecting the platform to Claude Code, Cursor, and VS Code Copilot. The platform is offered with HIPAA, SOC, and ISO certifications and supports cloud, self-hosted (Docker/Kubernetes), and customer-hosted deployment.

Comet serves two distinct buyer personas. Its traditional base is enterprise ML and AI engineering teams (data scientists, ML engineers, research teams) who need reproducible experiment tracking, model governance, and production monitoring. Its newer growth segment is LLM and GenAI application developers building chatbots, RAG pipelines, and agents who need observability, evaluation, prompt optimization, and — increasingly — cost governance over AI coding agents. The company sells through a hybrid go-to-market: a credit-card-free freemium tier for self-serve developers, quote-based enterprise contracts via direct sales, and a flagship distribution channel through Amazon SageMaker Partner AI Apps. Revenue mechanics combine annual subscription fees, usage-based consumption metered via Span telemetry and User Count, prepaid Comet Platform Credits for AI features (12-month expiry), and professional services for customer-hosted deployments. The company claims adoption by over 150,000 developers and 10,000+ teams, with named enterprise customers including Uber, Netflix, Google AI, Autodesk, Ancestry, Stellantis, Mobileye, NatWest, Zappos, Hugging Face, Stability AI, and AssemblyAI.

Comet ML, Inc. is a privately held, venture-backed company headquartered at 228 Park Ave S, Suite 15549, New York, NY. It has raised approximately $69.9 million in disclosed funding through a November 2021 Series B led by OpenView, with prior rounds including Scale Venture Partners, Trilogy Equity Partners, and Two Sigma Ventures. The company was founded in 2017 by Gideon Mendels (CEO) and Nimrod Lahav (CTO), acquired Stakion in April 2021, and was recognized as an Emerging Specialist in the 2025 Gartner Innovation Guide for Generative AI Engineering.

Name

Comet

Legal name

Comet ML, Inc.

Website

https://comet.com

Company type

Private

Founded year

2017

Operating status

Operating

Headcount range

11–50 employees

Short description

Comet ML, Inc. operates an AI developer platform combining classical MLOps (experiment tracking, model registry, production monitoring) with Opik, an open-source LLM observability and evaluation platform serving ML engineers and GenAI application developers through a freemium, enterprise, and AWS SageMaker distribution model.

Ownership category

akta.pro rank

Product category

AI Engineering Platform (MLOps & LLMOps)

NAICS

Computer Systems Design and Related Services (54151)

akta.pro primary industry

End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management) (HDAEANAA)

akta.pro secondary industries

AI Observability, Monitoring & Evaluation Platforms (Drift, Quality, Safety) (HDAEANAF), Enterprise AI Data & Knowledge Platforms (Vector Databases, Knowledge Graphs) (HDAEANAH), iBPMS (Intelligent BPM Suites) & Process Automation Suites (HDAEAHAG), AI/ML Solution Integration & MLOps Enablement (BPAEAAAJ), AI/ML Engineering & Model Development Services (BPAEAHAF)

Keywords

• ML experiment tracking
• LLM observability platform
• AI model evaluation
• MLOps infrastructure
• AI agent monitoring

Headquarters

HQ city

New York

HQ country

United States

HQ region

North America

Offices2 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Personnel, Technology or R&D, Infrastructure, Marketing or Sales, Operations

Revenue model

1. Platform Subscription Fees: Recurring subscription fees granting customers access to the Comet Platform during the Order Form Term, sold on annual/multi-year terms under a non-exclusive, non-transferable license.
2. Usage-Based Consumption (Span Usage & User Count): Consumption fees based on the number of Spans generated/processed/recorded by the Comet Platform and on User Count during the billing period, with overage billing at Order Form rates and Reasonable Usage protections.
3. Comet Platform Credits for AI Features: Pre-purchased, usage-based credits consumed when customers use AI Features (e.g., Opik Cost Intelligence, Agent Optimizer, AI assistant); credits expire within 12 months or at Order Form Term end, are non-refundable and non-transferable, and consumption rates can change.
4. Freemium / Free Tier: No-credit-card free tier for both Comet and Opik Cloud that users can keep using indefinitely, serving as the top-of-funnel land motion that converts into paid subscriptions and enterprise contracts.
5. Professional Services: Installation, configuration, and other professional services identified on an Order Form or SOW, including customer-hosted setup and reimbursed travel/out-of-pocket expenses.

Pricing tiers

Go-to-market motion4 records

Distribution channels5 records

Marketing channels10 records

Core offering

Comet operates the Comet AI Developer Platform, combining a hosted MLOps stack (experiment tracking, model registry, artifacts/dataset management, production monitoring) with Opik, an open-source LLM observability and evaluation platform. Opik records traces and tool/agent steps, scores outputs with 30+ LLM-as-a-judge and heuristic metrics, optimizes prompts via six algorithms, runs online evaluation rules, and includes modules for AI coding agent cost tracking (Cost Intelligence) and autonomous code fixes (Ollie).

Product overview

Comet operates a two-pronged product portfolio built on the same enterprise-grade platform: the legacy Comet ML MLOps suite (ML Experiment Management, ML Artifacts & Dataset Management, ML Model Registry, and Model Production Monitoring) for the full classical ML lifecycle, and Opik, the newer open-source AI engineering platform for LLM applications. Opik is the centerpiece and is delivered as a unified platform-plus-modules architecture that combines tracing/observability, annotation, evaluation (with 30+ LLM-as-a-judge metrics and test suites), prompt management with version control, and the Prompt Playground. On top of that core, Opik layers several distinct modules: Ollie (an autonomous coding agent that writes fixes back into the user's agent codebase), the Opik Agent Optimizer (six automated prompt/agent optimization algorithms), Cost Intelligence (Claude Code and Codex spend tracking), and an MCP Server (for connecting Claude Code, Cursor, or VS Code Copilot to the workspace). The MLOps suite remains a standalone product line tied to traditional ML frameworks (PyTorch, TensorFlow, Keras, Hugging Face, scikit-learn, XGBoost) and is offered on the same Comet infrastructure as Opik, including a tight integration with Amazon SageMaker Partner AI Apps.

Differentiator

Problem solved

Functional benefit

Brands

• Opik: Open-source LLM observability, evaluation, prompt optimization, and AI coding-agent cost tracking platform developed and maintained by Comet ML, Inc.

Products and services

• Opik Open-source LLM observability and evaluation platform that logs every LLM call, tool invocation, and agent step; supports annotation, dataset-based evaluation with 30+ LLM-as-a-judge and heuristic metrics, online evaluation rules, and production monitoring for GenAI applications.
• Opik Cost Intelligence Feature within Opik that gives engineering leaders real-time visibility into Claude Code and Codex coding-agent token spend per engineer, team, and task, surfacing waste from unused skills, idle MCPs, and misconfigured compaction strategies, with one-click optimizations.
• Ollie Built-in coding agent inside Opik that analyzes traces and test outcomes, identifies fixes, and writes them directly into the user's own agent's codebase, with version control and regression testing.
• Opik Agent Optimizer Optimization suite within Opik that automatically generates and tests prompt and tool variants for every step in an agent using six optimization algorithms (MetaPrompt, HRPO, Few-Shot Bayesian, Evolutionary, GEPA, Parameter, Tool Optimization) against user-defined metrics and datasets.
• ML Experiment Management Comet's core MLOps product for tracking and comparing machine learning model training runs with custom visualizations, model versioning, dataset and artifact management, and production monitoring, supported by SDK integrations with PyTorch, TensorFlow, Keras, Hugging Face, scikit-learn, and XGBoost.
• Model Production Monitoring (MPM) Purpose-built model monitoring module that tracks and visualizes model performance in production, detecting data drift and accuracy/error metrics, defining custom performance metrics, and triggering real-time alerts via configurable alerting rules; supports on-prem, VPC, and multi-cloud deployment.
• ML Artifacts & Dataset Management Dataset versioning and artifact management module of the ML Experiment Management platform for tracking and reproducing datasets, code, hyperparameters, and model artifacts across experiments.
• ML Model Registry Model versioning component of the ML Experiment Management platform that registers and versions trained models for reproducibility and governance across the model lifecycle.

Quantifiable outcome

• Average of 30% recoverable spend recovered from day one with Opik Cost Intelligence
• +5 more outcomes

Named customers14 records

Segments5 records

Ideal customer profiles5 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration13 records

AI capability13 records

Feature8 records

Partnerships

Four partnerships are on record, tiered flagship integration, core ai infrastructure partner, minor / adjacent (crypto/blockchain use case) and flagship / core distribution partnership.

• Amazon Web Services (AWS) — SageMaker integrationflagship integrationTechnology or Integration · 1 September 2025AWS published additional joint technical content ("AWS Integrates Comet into SageMaker for Scalable ML Experimentation") describing automatic logging and real-time insights for scalable ML experimentation with Comet's experiment management platform, addressing governance and compliance challenges for enterprises.
• OpenAIcore ai infrastructure partnerTechnology or Integration · 1 July 2025OpenAI is the generative-AI inference provider behind Comet's in-product AI-powered assistant feature. Per the Privacy Policy (last updated July 31, 2025), when users interact with the assistant, queries (potentially including personal data) are transmitted to OpenAI to generate responses; OpenAI processes the data only to provide the service and does not retain or use it for model training under its API data usage policy.
• Alita AIminor / adjacent (crypto/blockchain use case)Strategic or Co-development Partner · 1 March 2025Per CoinTrust coverage, Alita AI teamed up with "Comet" to streamline management of digital assets across Ethereum, Bitcoin, and other blockchains, employing AI to simplify blockchain transactions without needing code. The partnership aims to improve accessibility to decentralized finance and the crypto user experience. (Note: this relationship references the CometAPI/Comet brand in the blockchain space rather than Comet ML's AI engineering platform, and is included here as an adjacent brand collision flagged in the source.)
• Amazon Web Services (AWS)flagship / core distribution partnershipTechnology or Integration · 1 December 2024Comet integrated its AI model evaluation and experiment-tracking platform into Amazon SageMaker's new Partner AI Apps program, enabling joint customers to access Comet's tools for experiment tracking, LLM evaluation, and production monitoring directly inside the SageMaker environment. AWS also published a technical guide (Rapid ML experimentation for enterprises with Amazon SageMaker AI and Comet) demonstrating a fraud-detection workflow with reproducible, audit-ready logging. Announced via Business Wire and replicated across Comet, AWS, and trade-press coverage in late 2024 through 2025.

Scale indicators10 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• Neptune.ai: Experiment tracking and model registry platform for ML and LLM teams, with metadata store, model monitoring, and collaboration features that overlap with Comet's MLOps product line.
• WhyLabs: AI observability platform offering LLM and ML model monitoring, data drift, and guardrails, overlapping with Comet's Model Production Monitoring and Opik's online evaluation features.
• Weights & Biases: Closest direct competitor in MLOps (experiment tracking, model registry, sweeps) that has expanded into LLM evaluation and agent observability, serving the same ML/AI engineering personas as Comet.
• MLflow (Databricks): Open-source MLOps platform for experiment tracking, model registry, and deployment. Now owned by Databricks, it competes head-on with Comet's classical MLOps suite and has added LLM evaluation capabilities.
• LangSmith (LangChain): LLM-focused observability, evaluation, and debugging tool tightly integrated with LangChain. Direct overlap with Opik's tracing and evaluation for LLM applications and agents.
• Arize AI: Provides LLM observability, evaluation, and AI/ML monitoring. Competes directly with Opik on LLM tracing, online evaluation, and production drift detection for AI applications.

Emerging players

• Helicone: Open-source LLM observability platform with a focus on cost and usage tracking for LLM APIs. Overlaps with Opik's tracing and Cost Intelligence capabilities for AI engineering teams.
• Galileo (Galileo AI / Rungalileo): LLM evaluation and observability platform for AI applications and agents. Direct overlap with Opik's evaluation metrics, hallucination detection, and online evaluation rules.

Broad incumbents

• Datadog (with LLM Observability / Watchdog): Large observability and monitoring incumbent that has launched LLM Observability and AI monitoring products. Competes with Comet in enterprise production monitoring and as a unified telemetry platform.
• Dynatrace: Enterprise observability and APM incumbent expanding into AI/ML monitoring and AIOps. Competes with Comet for large-enterprise monitoring and governance budgets, particularly in regulated industries.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat6 records

Key risks6 records

Key highlights7 records

Customer concentration

• Xx.com/cometml
• GitHubgithub.com/comet-ml/opik
• YouTubeyoutube.com/channel/UCmN63HKvfXSCS-UwVwmK8Hw/featured
• LinkedInlinkedin.com/company/comet-ml/?viewAsMember=true

Compliance5 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles4 records

Subsidiaries1 record

Funding overview

Funding rounds5 records

Investors11 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A1 record

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Comet