100+ Free SnowPro Gen AI Practice Questions

Pass your SnowPro Specialty: Gen AI (GES-C01) exam on the first try — instant access, no signup required.

✓ No registration✓ No credit card✓ No hidden fees✓ Start practicing immediately

Not published Pass Rate

100+ Questions

100% Free

1 / 100

Question 1

Score: 0/0

Which Snowflake feature is the primary entry point for invoking large language model functions directly from SQL?

Snowpark Python UDFs

Snowflake Cortex

Snowflake Streams

Tasks and Procedures

to track

2026 Statistics

Key Facts: SnowPro Gen AI Exam

Not published

Pass Rate

Snowflake does not publish

750/1000

Passing Score

Scaled

60-100 hrs

Study Time

Recommended for prepared candidates

115 min

Exam Duration

Snowflake

$225

Exam Fee

Snowflake

2 years

Cert Valid

Snowflake recertification cycle

SnowPro Specialty: Gen AI (GES-C01) requires 750 out of 1000 to pass and runs 115 minutes. Domain weights: Snowflake for Gen AI Overview 26%, Snowflake Gen AI & LLM Functions 40%, Snowflake Gen AI Governance 22%, and Snowflake Document AI 12%. Prerequisites include current SnowPro Core or SnowPro Associate Platform plus 1+ years of Gen AI experience and Python/SQL/Data Engineering proficiency. The exam fee is $225 USD.

Sample SnowPro Gen AI Practice Questions

Try these sample questions to test your SnowPro Gen AI exam readiness. Each question includes a detailed explanation. Start the interactive quiz above for the full 100+ question experience with AI tutoring.

1Which Snowflake feature is the primary entry point for invoking large language model functions directly from SQL?

A.Snowpark Python UDFs

B.Snowflake Cortex

C.Snowflake Streams

D.Tasks and Procedures

Explanation: Snowflake Cortex is the platform's built-in suite of managed AI services that exposes LLM capabilities (COMPLETE, SUMMARIZE, TRANSLATE, EMBED_TEXT_*, CLASSIFY_TEXT, EXTRACT_ANSWER, SENTIMENT, and AI_* aggregate functions) as SQL functions. Users do not need to provision or manage the underlying GPUs, models, or APIs.

2Which of the following is a foundation model hosted natively on Snowflake Cortex?

A.Snowflake Arctic

B.GPT-4o

C.Gemini 1.5 Pro

D.Claude 3 Opus

Explanation: Snowflake Arctic is Snowflake's own enterprise-focused foundation model, hosted natively on Cortex alongside other supported models such as Llama 3.x, Mistral, Mixtral, Reka, and Jamba. GPT, Gemini, and Claude are not natively hosted on Cortex.

3How are Cortex LLM Functions billed in Snowflake?

A.A fixed monthly subscription per user

B.Snowflake credits consumed proportional to the number of tokens processed

C.Per-second virtual warehouse runtime only

D.Free of charge for any account on Enterprise edition or above

Explanation: Cortex LLM Functions follow a credits-per-token cost model. Each model has a published credits-per-million-tokens rate, and consumption is metered against your Snowflake account's credit balance, separate from warehouse compute charges.

4Which Snowflake data type is used to store fixed-length numerical arrays such as embeddings?

A.ARRAY

B.VARIANT

C.VECTOR

D.OBJECT

Explanation: The VECTOR data type stores fixed-length numeric arrays optimized for similarity search. It is parameterized by element type and dimension (for example VECTOR(FLOAT, 768)) and works with VECTOR_COSINE_SIMILARITY, VECTOR_INNER_PRODUCT, and VECTOR_L2_DISTANCE.

5What is the primary purpose of Snowflake Notebooks in a Gen AI workflow?

A.To replace virtual warehouses with serverless compute

B.To provide a unified, interactive Python and SQL development environment for AI and ML work inside Snowflake

C.To store fine-tuned LLM model weights

D.To act as a managed vector database service

Explanation: Snowflake Notebooks are an interactive development surface inside Snowsight that supports Python and SQL cells. They are commonly used to prototype Cortex calls, build RAG pipelines with Cortex Search, and iterate on Document AI extraction, all without leaving the platform.

6Which Snowflake service is purpose-built for retrieval-augmented generation (RAG) over enterprise text data?

A.Snowflake Streams

B.Cortex Search

C.Snowpipe

D.Dynamic Tables

Explanation: Cortex Search is a fully managed hybrid search service that combines vector similarity and keyword (BM25-style) retrieval to support RAG patterns. It indexes Snowflake-resident text and is commonly paired with COMPLETE for grounded answers.

7Which Snowflake feature enables conversational, natural-language analytics over structured data using a semantic model?

A.Cortex Analyst

B.Cortex Guardrails

C.Cortex Fine-Tuning

D.Snowflake Streamlit

Explanation: Cortex Analyst lets users ask natural-language questions over structured tables. It uses a YAML semantic model that defines tables, columns, metrics, and synonyms, and translates questions into SQL the user can review.

8Which Snowflake offering is most appropriate for serving a custom open-source LLM that is NOT in the Cortex catalog?

A.Cortex COMPLETE

B.Snowpark Container Services

C.External Functions

D.Snowflake Search Optimization Service

Explanation: Snowpark Container Services lets you run long-running container workloads, including GPU-backed services, inside Snowflake. It is the recommended option for serving custom LLMs or open-source models that Cortex does not host natively, while keeping inference next to governed data.

9Which statement BEST describes how data is handled when calling a Cortex LLM Function in Snowflake?

A.Data is sent to a public internet API for processing

B.Inference runs inside Snowflake's governance boundary; prompt and response are not used to train the foundation model

C.Data is automatically copied to a third-party vector database

D.Cortex requires the customer to provide their own OpenAI API key

Explanation: Cortex LLM Functions execute inside the Snowflake account's governance boundary. Snowflake states that customer prompts and responses are not used to train the underlying foundation models, which is a key reason Cortex is positioned for regulated enterprise workloads.

10Which of the following best describes Snowflake Arctic Embed?

A.A SQL function that fine-tunes Llama models

B.A family of Snowflake-developed text embedding models exposed through Cortex EMBED_TEXT_768 and EMBED_TEXT_1024

C.A vector database product separate from Snowflake

D.A feature store for Snowpark ML

Explanation: Snowflake Arctic Embed is Snowflake's family of open text embedding models. Cortex exposes them through EMBED_TEXT_768 and EMBED_TEXT_1024, producing 768- or 1024-dimension VECTOR outputs typically used for similarity search and Cortex Search indexes.

About the SnowPro Gen AI Exam

The SnowPro Specialty: Gen AI (GES-C01) certification validates the ability to design, build, and govern generative AI applications on Snowflake. It covers Cortex LLM Functions (COMPLETE, SUMMARIZE, TRANSLATE, EMBED_TEXT_768/1024, CLASSIFY_TEXT, AI_FILTER, AI_AGG), VECTOR similarity search, Cortex Search, Cortex Analyst, Cortex Fine-Tuning, Cortex Guardrails, Document AI, and AI governance using masking, row access policies, tags, lineage, and the Trust Center.

Questions

65 scored questions

Time Limit

115 minutes

Passing Score

750/1000 (scaled)

Exam Fee

$225 USD (Snowflake Inc.)

SnowPro Gen AI Exam Content Outline

40%

Snowflake Gen AI & LLM Functions

Cortex COMPLETE, SUMMARIZE, TRANSLATE, SENTIMENT, EXTRACT_ANSWER, CLASSIFY_TEXT, EMBED_TEXT_768/1024, AI_FILTER, AI_CLASSIFY, AI_AGG, AI_SUMMARIZE_AGG, AI_SIMILARITY, VECTOR similarity, Cortex Search, Cortex Analyst, Cortex Fine-Tuning, RAG patterns

26%

Snowflake for Gen AI Overview

Snowflake architecture for AI, supported foundation models (Snowflake Arctic, Llama 3.x, Mistral, Mixtral, Reka, Jamba), Snowflake Arctic Embed, VECTOR data type, Snowflake Notebooks, Streamlit in Snowflake, Snowpark Container Services, cost model and credits per token

22%

Snowflake Gen AI Governance

Cortex Guardrails, dynamic data masking, row access policies, tags and data classification, model lineage, ACCESS_HISTORY auditing, the Trust Center, regional data residency, and least-privilege RBAC for Cortex

12%

Snowflake Document AI

Document AI Studio, defining text/numeric/checkbox fields, training and improving models on sample documents, running extraction at scale from stages, confidence-based routing, and combining Document AI with Cortex Search and COMPLETE

How to Pass the SnowPro Gen AI Exam

What You Need to Know

Passing score: 750/1000 (scaled)
Exam length: 65 questions
Time limit: 115 minutes
Exam fee: $225 USD

Keys to Passing

Complete 500+ practice questions
Score 80%+ consistently before scheduling
Focus on highest-weighted sections
Use our AI tutor for tough concepts

SnowPro Gen AI Study Tips from Top Performers

1Spend the most preparation time on Cortex LLM Functions (40%) — know each task function and when to choose COMPLETE vs SUMMARIZE/TRANSLATE/SENTIMENT/EXTRACT_ANSWER/CLASSIFY_TEXT

2Memorize the VECTOR similarity functions and their sort directions: cosine and inner product DESC for nearest, L2 distance ASC for nearest

3Practice the canonical RAG pattern: chunk, embed with EMBED_TEXT_768/1024, retrieve via Cortex Search, ground COMPLETE

4Understand AISQL aggregate functions (AI_FILTER, AI_AGG, AI_SUMMARIZE_AGG, AI_CLASSIFY) and when set-based beats per-row UDFs

5Map governance to AI: masking and row access policies on inputs, Cortex Guardrails on prompts/responses, tags + classification + lineage for accountability

6Know Document AI field types (text/numeric/checkbox) and the Studio iteration loop for accuracy improvements

7Be ready for cost-optimization questions: tokens-per-call, model tiering, embed-once persistence, region/edition availability

Frequently Asked Questions

What is the SnowPro Specialty: Gen AI passing score?

GES-C01 requires a passing score of 750 out of 1000 on a scaled scoring system. The exam runs 115 minutes. There is no penalty for incorrect answers, so candidates should attempt every question.

What are the prerequisites for the GES-C01 exam?

Snowflake recommends a current SnowPro Core or SnowPro Associate Platform certification, at least one year of Gen AI experience, and proficiency in Python, SQL, and data engineering. There is no degree requirement.

What domains are on the SnowPro Gen AI exam?

Four domains: Snowflake for Gen AI Overview (26%), Snowflake Gen AI & LLM Functions (40%), Snowflake Gen AI Governance (22%), and Snowflake Document AI (12%). The largest domain is Cortex LLM Functions, including COMPLETE, EMBED_TEXT_*, AI_FILTER/AGG/CLASSIFY, Cortex Search, and Cortex Analyst.

How much does GES-C01 cost?

The exam fee is $225 USD. Retakes require full payment per attempt and a minimum 7-day waiting period between attempts.

Which foundation models are hosted on Snowflake Cortex?

Cortex hosts a curated catalog including Snowflake Arctic, Llama 3.x, Mistral and Mixtral, Reka, and Jamba, plus the Snowflake Arctic Embed family for embeddings (EMBED_TEXT_768 and EMBED_TEXT_1024). Models from third parties such as GPT, Claude, and Gemini are not natively hosted in Cortex.

How is Cortex billed?

Cortex LLM Functions are billed in Snowflake credits proportional to the tokens processed by each model, separate from warehouse compute charges. Cost optimization usually focuses on choosing the right model for each task and avoiding unnecessary re-embedding.

Is Document AI a separate product?

Document AI is a Snowflake-managed capability accessed through Document AI Studio. It extracts text, numeric, and checkbox fields from PDFs and images, returns per-field confidence scores, and is governed by the same RBAC, masking, and lineage as the rest of Snowflake.