Snowflake Cortex AI 2026: Features, Pricing & Uses

Most companies want AI and ML, but setting up the infrastructure, moving data around, and hiring specialists can be painful. Snowflake Cortex AI fixes that. It’s a fully managed AI layer built right into Snowflake. No separate clusters, no pipelines, and your data stays put. Run LLMs, agents, and predictions using plain SQL on data you already have.

In April 2026, Snowflake positioned Cortex as the control plane for the “agentic enterprise,” with big updates to Snowflake Intelligence, Cortex Code, and its $200M OpenAI partnership. This guide reflects the current state of Snowflake Cortex AI.

What Is Snowflake Cortex AI?

Snowflake Cortex AI is Snowflake’s built-in AI and machine learning service. It covers LLM-powered analytics, semantic search, smart agents, and ML functions. Your data doesn’t need to go anywhere. No GPU setup to deal with. No Python from scratch either.

Snowflake rolled this out back in late 2023. Started small, just a few SQL-callable AI functions. But by 2026, it’s turned into a full-blown AI platform. If you’re evaluating where it fits in your data strategy, working with an experienced Snowflake consulting partner can help you prioritize the right capabilities from the start.

AI SQL functions (text analysis, generation, embeddings)
Cortex Search (hybrid semantic + keyword retrieval)
Cortex Agents (multi-step autonomous workflows)
Cortex Analyst (natural language to SQL)
Snowflake Intelligence (enterprise-wide AI agent layer)
Cortex Code (AI-assisted data engineering and app development)
Cortex Fine-tuning (domain-specific model customization)
Document AI / AI_EXTRACT (structured extraction from unstructured content).

Planning an AI initiative on Snowflake?

Our Snowflake AI experts help enterprises evaluate use cases, select the right Cortex capabilities, and build secure AI solutions without moving data outside Snowflake.

Talk to our Snowflake AI Consultants

Key Updates on Snowflake Cortex AI in 2026

$200M OpenAI Partnership (February 2026): OpenAI frontier models natively embedded across AWS, Azure, and GCP inside Cortex AI
AI Credits Overhaul (April 2026): Flat $2.00/credit globally, decoupled from Snowflake Edition, up to 70% cost reductions reported
Snowflake Intelligence (GA): Agentic Document Analytics aggregates insights across thousands of documents, beyond traditional RAG
Cortex Code CLI: Now supports dbt and Apache Airflow; used by 50%+ of Snowflake customers
AI_EXTRACT (GA October 2025): Structured extraction from unstructured documents, generally available
Gemini 3.5 Flash: Multimodal video and audio analysis now available via AI_COMPLETE
AI_TRANSCRIBE pricing: Reduced by up to 60% effective June 1, 2026

How Cortex AI Differs from Traditional AI?

Normally, building an AI stack means stitching a bunch of things together. Data pipelines, vector databases, model serving setups, and a whole engineering team to keep it running. Cortex AI just skips all that.

Factor	Traditional AI	Snowflake Cortex AI (2026)
Infrastructure	Separate GPU clusters	Fully managed, serverless
Data Movement	Data is sent to external systems	Stays inside Snowflake perimeter
Time to Deploy	Weeks to months	Hours to days via SQL
Skill Required	Data scientists + ML engineers	SQL analysts can start immediately
Cost Model	Fixed infra + specialist salaries	AI Credits, token-based pay-per-use
Model Access	Limited to deployed models	Anthropic, OpenAI, Meta, Mistral, Google
Agentic Capability	Custom-built per integration	Snowflake Intelligence + MCP (GA 2026)

How Does Snowflake Cortex AI Work?

Snowflake Cortex AI brings generative AI directly into your Snowflake environment, eliminating the need for separate AI infrastructure or data movement. Teams can access AI capabilities using SQL while keeping their data within Snowflake’s security and governance framework.

For example, analysts can run AI functions directly from SQL:

SELECT SNOWFLAKE.CORTEX.SENTIMENT(customer_feedback)

FROM support_tickets;

Behind the scenes, Snowflake routes requests to foundation models from providers like OpenAI, Anthropic, Google, Meta, and Mistral. Execution is fully serverless, automatically scaling based on workload while preserving existing controls such as RBAC, masking policies, and audit logs.

Snowflake Data → Cortex AI Layer → Foundation Models → AI Output

The result is faster AI adoption, enabling organizations to build search, analytics, and AI applications directly where their data already lives.

Snowflake Cortex AI Stack Core Features

1. Cortex AI Functions (AISQL)

The most accessible entry point. AISQL lets analysts invoke AI directly inside SQL queries with no Python or ML framework required.

Snowflake Cortex AI key functions include:

AI_COMPLETE — LLM access for generation, reasoning, and custom prompts
AI_EXTRACT — Structured field extraction from unstructured documents (GA October 2025)
SENTIMENT — Text scoring from -1 (negative) to +1 (positive)
SUMMARIZE — Automatic long-form content condensing
EMBED_TEXT — Vector embeddings for similarity search and RAG
TRANSLATE — Text conversion across 60+ languages
ANALYST — Natural language to optimized SQL

sql
SELECT AI_COMPLETE(
  'claude-sonnet-4-6',
  CONCAT('Summarize this support ticket: ', ticket_body)
) AS summary
FROM support_tickets
WHERE created_date >= CURRENT_DATE - 7;

AI Function Studio (Public Preview, 2026) automates prompt engineering, model selection, and benchmarking. This reduces experimentation overhead before scaling to production.

2. Cortex Search

Cortex Search combines semantic vector search with keyword matching in a single managed service. No separate vector database, no ETL. It handles embedding, indexing, and retrieval automatically and is the backbone of RAG applications inside Snowflake.

Note: Cortex Search carries an always-on serving charge per GB indexed suspend dev services when idle to avoid unnecessary costs.

3. Cortex Agents

Cortex Agents handle multi-step autonomous workflows across structured and unstructured data. They break complex questions into components, route sub-tasks to the appropriate model, and return grounded, source-aware responses.

Key 2026 extensions: native Microsoft Teams and Copilot integration (Preview), MCP Server for connecting to Salesforce Agentforce, UiPath, and Anthropic platforms, and Snowflake Intelligence for analytical queries across thousands of documents simultaneously.

4. Cortex Analyst, Fine-tuning & Cortex Code

Cortex Analyst translates natural language into optimized SQL, bundled at no extra cost for Enterprise Edition customers. Cortex Fine-tuning adapts foundation models to domain-specific terminology without leaving Snowflake, reducing token usage over time.

Cortex Code, launched in November 2025, provides AI-assisted development via Snowsight or CLI, now supporting dbt and Apache Airflow alongside Claude Opus 4.6 and GPT-5.2.

Not sure which Cortex capability fits your use case?

Whether you're building RAG applications with Cortex Search, deploying AI agents, or enabling text-to-SQL analytics, our team can help you design the right Snowflake architecture.

Schedule a Snowflake Cortex AI Assessment

Benefits of Snowflake Cortex AI for Enterprises

The biggest advantage of Snowflake Cortex AI isn’t just access to large language models. It’s the ability to bring AI directly to where your data already lives.

Benefit	Description
No Data Movement	AI runs where data lives
Faster Deployment	Deploy AI in hours instead of months
Lower Operational Overhead	No GPUs or vector DBs to manage
Enterprise Governance	Existing RBAC and masking apply
Access to Multiple LLMs	OpenAI, Anthropic, Gemini, Meta
Cost Efficiency	Pay only for usage

Snowflake Cortex AI Pricing 2026

Starting April 1, 2026, Snowflake added something new called AI credits, basically a separate billing currency. Costs $2.00 per credit globally, or $2.20 if you’re on a regional plan. And here’s the thing: Standard and Enterprise customers now pay the same rate for AI inference. No tier difference there anymore.

Cost Structure

Cortex AI billing covers four categories: AI token charges (per million tokens, input and output billed separately), Cortex Search serving (always-on, per GB/month), Document AI processing (per document), and standard warehouse compute.

Model Tier	Input Rate	Output Rate	Best For
Small (LLaMA 3.1–8B, Mistral-7B)	~$0.12–$0.20/M	~$0.12–$0.20/M	Classification, routing, bulk summarization
Mid-range (Arctic, Mistral Large)	~$0.50–$1.00/M	~$1.50–$3.00/M	Balanced reasoning tasks
Frontier (Claude Sonnet 4-6, GPT-4o)	~$1.50–$3.00/M	~$5.00–$10.00/M	Complex reasoning, agents

Real-world check:

A single AI_COMPLETE call across one million rows on a frontier model can exceed 26,000 AI Credits.
The same workload on a small open-source model: ~440 credits.
Model selection is your single largest cost lever.
Prompt caching reduces repeated-context costs by ~90%.
Teams combining caching with smart model routing report 70%+ total cost reductions vs. pre-April baselines.
Track spend via CORTEX_INFERENCE_USAGE_HISTORY and CORTEX_AI_PRICING and build custom alerts, since no native AI Credit resource monitors exist.

Concerned about AI costs?

We help organizations optimize AI Credits, implement prompt caching, and choose the right models to reduce Snowflake Cortex AI spend while maintaining performance.

Explore Snowflake Cost Optimization Services

Snowflake Cortex AI Use Cases Across Industries

Customer Experience: Real-time sentiment scoring on tickets and chat logs, LLM-based ticket routing, and personalized response generation at scale.
Financial Services: Back in October 2025, Snowflake rolled out a Cortex AI package built specifically for financial services. Covers stuff like fraud detection, going through compliance documents, and catching weird trading patterns. There are MCP-connected agents in there too, hooking into Salesforce and UiPath. And since the data never leaves Snowflake, it kind of takes care of GDPR, HIPAA, and SOC 2 compliance, all in one go.
Document Intelligence: Structured extraction via AI_EXTRACT, enterprise RAG chatbots via Cortex Search + Agents, and Agentic Document Analytics for aggregating insights across thousands of documents simultaneously. Use AI_EXTRACT over full SUMMARIZE for targeted contract queries, significantly cheaper per token.
Predictive Analytics: Time-series forecasting, anomaly detection, and contribution analysis via Cortex ML functions, without external ML infrastructure.

Cortex AI vs. Databricks vs. Azure OpenAI

Dimension	Snowflake Cortex AI	Databricks Mosaic AI	Azure OpenAI
Primary Strength	SQL-first AI on a governed warehouse	Full ML lifecycle, custom model training	Direct OpenAI API access
Skill Floor	SQL analysts	Data scientists + ML engineers	API developers
Model Training	Fine-tuning only	Full training (MLflow, GPU clusters)	None (inference only)
Governance	Native Snowflake RBAC	Unity Catalog	Azure IAM
Multi-cloud	AWS, Azure, GCP	AWS, Azure, GCP	Azure only
Best Fit	SQL teams on Snowflake	ML-heavy teams building custom models	Azure-native orgs

Cortex AI wins on time-to-value and governance for regulated industries. Databricks services win for custom model training and complex MLOps. Azure OpenAI is the right fit if you’re deeply embedded in Microsoft’s ecosystem.

Limitations & When Not to Use Cortex AI

There are times when Cortex just isn’t the right fit. Like if you need to train foundation models from scratch, Cortex only does fine-tuning, not that. Or if you need millisecond-level edge inference, the serverless setup adds latency, so that’s out too.

The same goes if your data doesn’t live in Snowflake and migrating isn’t really an option right now. And if you’re after full MLOps pipelines with experiment tracking, honestly, Databricks Mosaic AI handles that better.

Honest limitation: For most companies, the platform usually isn’t the bottleneck. It’s the data quality. That’s where the real work is. Poorly modeled or ungoverned Snowflake data means fast deployment just produces fast bad outputs.

How to Get Started with Snowflake Cortex AI?

Phase 1 – Assessment: During the first two weeks of Snowflake implementation, your goal is simply to understand where Cortex AI fits. Identify 2–3 text-heavy use cases with measurable ROI. Review governance posture. Run sample token tests to baseline costs. Define success metrics before touching production data.

Phase 2 – Pilot: Start with SENTIMENT, SUMMARIZE, or AI_EXTRACT for immediate value without full agent orchestration. Tag all AI queries for cost attribution, monitor via CORTEX_INFERENCE_USAGE_HISTORY, and validate outputs against human review before automating decisions.

Phase 3 – Scale: Introduce Cortex Agents for multi-step workflows, implement prompt caching, and consider fine-tuning for recurring domain patterns. Formalize model evaluation and human-in-the-loop requirements before full production rollout.

Accelerate Enterprise AI with Snowflake Cortex AI

From semantic search and AI agents to text-to-SQL and document intelligence, our Snowflake experts help enterprises design, deploy, and optimize Cortex AI solutions that deliver measurable business outcomes.

Our services include:

Snowflake Cortex AI implementation
AI architecture and governance
Semantic model design
RAG and agent development
Cost optimization and monitoring
Enterprise AI consulting

Ready to unlock AI on your Snowflake data?

Book a Free Snowflake Cortex AI Consultation

Frequently Asked Questions

What is the difference between Snowflake and Snowflake Cortex?

Snowflake is a cloud data platform used for storing, processing, and analyzing data. Snowflake Cortex is the AI layer within Snowflake that provides capabilities such as text-to-SQL, semantic search, AI agents, and LLM-powered analytics. In simple terms, Snowflake manages data, while Snowflake Cortex enables AI-driven insights and automation on that data.

Which LLMs does Snowflake Cortex support?

Anthropic (Claude Sonnet 4-6, Opus 4.6), OpenAI (GPT-4o, GPT-5.2), Meta (LLaMA 3.1 8B/70B), Mistral, Google (Gemini 3.5 Flash), and Snowflake Arctic are updated continuously.

Is Cortex AI secure and compliant?

Yes. All inference runs inside your Snowflake cloud region. Existing RBAC, column masking, row-level security, and audit logs apply to all AI outputs.

Can I use Cortex AI without a data science background?

Yes. AISQL requires only SQL. Cortex Analyst accepts natural language. Snowflake Intelligence provides a no-code interface. Fine-tuning benefits from Python expertise.

What’s the difference between Cortex Agents and Snowflake Intelligence?

Cortex Agents is the developer platform for programmatic multi-step workflows. Snowflake Intelligence is the business-user layer on top. No-code interfaces, cross-document analytics, and native Jira/Salesforce integrations.

Is Snowflake Cortex AI serverless?

Yes. Most Cortex AI capabilities are fully managed and serverless, allowing organizations to use AI services without provisioning infrastructure or managing GPU clusters.

Need help accelerating your pilot?

Our team helps enterprises launch Snowflake Cortex AI use cases faster with governance, model evaluation, and cost controls built in.

Get Started with Snowflake Cortex AI

1. Cortex-AI Overview

Snowflake Gen-AI/ML capabilities is empowered by “Cortex-AI”, this is a framework under which there are a lot of features which are available and each of them are being used for a specific use case around solving AI, Gen-AI and ML workloads. Pictorially the entire suit of Cortex-AI looks like below:

2. Cortex-AI features

Now, let us understand the Cortex-AI features and its capabilities on a high level.

Cortex-LLM functions: Snowflake Cortex LLM function provides seamless access to top-tier large language models (LLMs) developed by leading researchers at companies such as Anthropic, Mistral, Reka AI, Meta, and Google. This includes Snowflake Arctic, an enterprise-grade open model created by Snowflake. Since these LLMs are fully hosted and managed within Snowflake, no setup is required. Your data remains securely within Snowflake, ensuring the performance, scalability, and governance you rely on.
Cortex Analyst: Cortex Analyst is a fully managed Snowflake Cortex feature that enables applications to deliver accurate answers to business queries using structured data. With natural language support, users can obtain insights without writing SQL. Available via a REST API, it seamlessly integrates into any application. Henceforth with this feature taking form and shape business users would be able to write the SQLs on structured data with plain and simple English.
Document AI: Document AI, a Snowflake AI feature, leverages Arctic-TILT, a proprietary large language model (LLM), to extract data from various document formats. It processes both text-heavy content and graphical elements like logos, handwritten signatures, and checkmarks. Designed for automated workflows, Document AI enables seamless pipeline creation for continuously processing specific document types, such as invoices or financial statements be it in form of PDF, word document, etc..
Cortex Search: Cortex Search provides low-latency, high-quality fuzzy search across your Snowflake data, supporting diverse search experiences, including Retrieval Augmented Generation (RAG) applications powered by Large Language Models (LLMs).With a hybrid search engine that combines vector and keyword search, Cortex Search allows you to quickly set up text-based search capabilities in minutes—without the need for manual embedding, infrastructure management, search tuning, or index maintenance.
Cortex Agents: Cortex Agents coordinate across structured and unstructured data sources to generate insights. They plan, execute tasks using specialized tools, and produce responses. Leveraging Cortex Analyst for structured data and Cortex Search for unstructured data, along with LLMs, Cortex Agents enable comprehensive data analysis. Cortex Search extracts insights from unstructured sources, while Cortex Analyst translates queries into SQL for structured data processing.

In this blog, we would primarily focus on Cortex-LLM functions which are fully hosted and managed by Snowflake.

3. Cortex LLM functions

Cortex LLM functions are available within Snowflake as SQL functions and also available in Python. Primarily these can be broadly categorized into the below set of functions.

All these Cortex-LLM functions are empowered by the usage of the Large Language models of various providers like Mistral, Meta, Anthropic, etc. Hence the usage of these models would drive the sizing and cost. The bigger the model size, the higher no, of tokens was used to train it and hence it is expected to be cost more. Below is the pictorial depiction on how the models and the categorization are done within the platform.

This is how the models have been used and its categories based on the use case, the same can be used and also improve the accuracy of the results.

Now, let us understand some of the key LLM functions and how they work within the platform.

A. COMPLETE:

Let us now see how the COMPLETE function works within the AI data cloud platform.

Cortex-Complete function with various models

If we see this we can ask very simple questions and start getting answers by the usage of various LLM models. The above image shows how in each of the question a separate model is used and it can give the response accordingly.

Input & Output of Cortex Complete functions

The above is an example we use the Cortex-Complete via SQL and the same generates a response by answering the question.

B. SUMMARIZE (task specific functions):

The SUMMARIZE function generates a concise summary of the provided English text. It is useful for condensing large paragraphs into a one- or two-line summary, making it easier to extract key information. Here’s an example:

C. TRANSLATE (task specific functions):

Snowflake Cortex LLM TRANSLATE functions enable text translation between multiple supported languages. This function allows seamless conversion of text from one language to another using the following supported language codes:

Now, let us see an example as how it works and it can convert the various languages to the language of choice.

As we see when the above function is used, it can seamlessly take the prompt as an input, recognizes it and then convert it as an output and that too in English.

D. SENTIMENT (task specific functions):

This is a function that does a sentiment analysis based on the input provided and gives a score between -1 and 1. The value “-1” being negative and “+1” being positive and “0” being neutral.

If we see this function, what it did was based on the prompt that we gave, it derived the scoring and because it was an optimistic statement hence the response was near to 1.

E. COUNT_TOKENS (helper functions):

This function calculates the number of tokens in a given prompt for a large language model or a specified task-specific function. It does not support fine-tuned models. This function can be a very good asset when it comes to evaluating the cost, as it gives us the information about the tokens getting scanned OR used so that the model gives the output.

In this example we see how the MIXTRAL LLM model was used to assess the prompt based on the number of tokens that prompt has. Like-wise there are multiple models we can use to ensure we get the value of the tokens.

To summarize, we have seen how Cortex COMPLETE, TASK SPECIFIC functions and helper functions work within the platform. These are natively available within the platform so that various workloads can benefit from integrated AI capabilities. Expert Snowflake development services further help in automating, scaling, and customizing these functions for industry-specific applications.

4. COST CONSIDERATIONS

Snowflake Cortex LLM functions offer a powerful way to extract insights from structured and unstructured data, but understanding their costs and usage limits is crucial. Compute costs are based on the number of tokens processed, where a token represents the smallest text unit, approximately four characters. The cost per million tokens varies depending on the specific LLM function and model used.

There are two important things which needs to be monitored. The first one is how do we track the AI Services cost. Now, this cost is a measure of all the AI services within the account which includes the Cortex-LLM functions. The key SQL which can be used are:

SELECT * FROM SNOWFLAKE.ACCOUNT_USAGE.METERING_DAILY_HISTORY
  WHERE SERVICE_TYPE='AI_SERVICES';

Another important point is to have a way to track just the usage of the Cortex LLM functions. Over here, we can refer the object name i.e.,

SNOWFLAKE.ACCOUNT_USAGE.CORTEX_FUNCTIONS_USAGE_HISTORY, it is where we get the exact credits consumed by the usage of various model in the LLM functions.

If we want to establish a relationship between the objects of the consumption i.e.,

“SNOWFLAKE.ACCOUNT_USAGE.METERING_DAILY_HISTORY” and “SNOWFLAKE.ACCOUNT_USAGE.CORTEX_FUNCTIONS_USAGE_HISTORY”

then we can in most of the cases see the o/p of the below 2 queries matching:

5. Date Engineering use case:

There are so many possibilities within this AI data cloud which can be solved by these LLM functions over here we would be demonstrating how we can derive quick insights from an attribute with very minimum hassle. With the right use case identification and model tuning, guided by Snowflake consulting professionals, even complex tasks like deriving missing attributes can be streamlined with LLM functions.

Let us say we have a CUSTOMER table, which has all the attributes but that doesn’t have a key column which is “Gender”, and that needs to be derived based on the data which is available in the table. Now, this is important because if we know these attributes we can easily do the customer segmentation, do advanced analytics and even we can send customized benefits. Let us consider we have a table named “DEMO_CUSTOMER_LLM”, which has the data sets as mentioned below:

Now, if there is an ask to get the Gender names by looking up to the Salutation field and imagine if this table is having millions of records, then how do we solve this problem? This is exactly where we can use the Cortex LLM functions and get the “gender” value. Please see the below section on how it is being used to get the Gender value.

Gender for most of the cases using Prompt engineering

If we see over here, for most of the cases by looking up to the SALUTATION field, we got the gender values, the model even has given us the suggestions whenever there is a generic title that has been used. That is how we can get quick insights using the LLM functions. We can even use this syntax and try multiple models to get the similar output.

Cortex-LLM functions enhance data engineering by enabling AI-driven automation, natural language interactions, and intelligent search. While they simplify AI adoption in Snowflake, organizations must optimize token usage to control costs effectively. It simplifies data processing, enabling easy access to complex queries and insights, even for non-technical users. This enhances efficiency and productivity in data operations.

in (infinity)
-inf (negative infinity)

These values are case-insensitive and should be enclosed in single quotes. It’s important to note that Snowflake handles ‘NaN’ differently from the IEEE 754 standard, treating all ‘NaN’ values as equal and considering ‘NaN’ to be greater than any other FLOAT value.

Summary

As we have seen Snowflake supports variety of data types that makes this platform unique with respect to handling various forms and formats of the data sets these features apart from many others makes this platform very unique for managing and analyzing the data at scale. Snowflake’s data types as a set of keys on a keyring, each designed to unlock a specific door to solve a particular data challenge. Whether it’s numeric precision, text processing, or managing complex JSON structures, each key represents a unique data type tailored for the task. Just like selecting the right key ensures seamless access, choosing the right data type in Snowflake unlocks the full potential of your data efficiently and effectively.

Post Tags

#Snowflake Cortex LLM

Yash Shah

Yash Shah is a seasoned Data Warehouse Consultant and Cloud Data Architect at Aegis Softtech, where he has spent over a decade designing and implementing enterprise-grade data solutions. With deep expertise in Snowflake, AWS, Azure, GCP, and the modern data stack, Yash helps organizations transform raw data into business-ready insights through robust data models, scalable architectures, and performance-tuned pipelines.He has led projects that streamlined ELT workflows, reduced operational overhead by 70%, and optimized cloud costs through effective resource monitoring. He owns and delivers technical proficiency and business acumen to every engagement.

Snowflake Cortex AI 2026: Complete Guide to Features, Pricing & Use Cases