GPT-4 in HighLevel

Modified on: Thu, 21 Aug, 2025 at 7:17 AM

This article explains how to use GPT-powered AI models, including the latest GPT-4 Turbo inside HighLevel Workflows. Learn how to set up AI steps, configure smart prompts, and automate powerful actions like lead follow-ups, conversation summaries, and personalized messaging.

Please Note: This article is about the Chat GPT and its version GPT-4. If you wish to learn about How to Use AI in Workflows, refer to Workflow Action - GPT Powered by OpenAI

TABLE OF CONTENTS

GPT-4 Turbo: Smarter AI with High-Level
What’s New
Pricing of GPT-4
Use Cases and Example Prompts
GPT Models Compared
The GPT-Generated Output
Frequently Asked Questions
Related Articles

GPT-4 Turbo: Smarter AI with High-Level

With the rise of conversational AI, businesses need tools that are smart, scalable, and deeply integrated. HighLevel now brings this power to your automation engine through GPT-powered AI Workflow steps. From writing personalized follow-ups and summarizing conversations to generating dynamic email content and social captions, you’ll learn how to make your automations more intelligent, contextual, and human-like with minimal setup.

Whether you're a marketing agency, service provider, or sales team, this guide will help you unlock powerful new capabilities using AI in your existing HighLevel workflows.

What’s New

GPT-4 Turbo: Now available in AI steps with faster response times and lower cost-per-call compared to previous GPT-4 models.
Model Switcher UI: Select between GPT-3.5, GPT-4, or GPT-4 Turbo directly in your workflow step.
Improved Prompt Handling: Enhanced merge field support and better handling of complex instructions.
System Prompt (optional): Add tone or behavior guidance to shape the AI’s personality and output style.

Pricing of GPT-4

HighLevel’s GPT Action is now billed at Zero Markup, meaning you pay exactly what OpenAI charges with no added fees. Pricing is token-based, calculated from the number of input tokens (your prompts) and output tokens (AI responses), and varies depending on the model selected (e.g., GPT-4o or GPT-4o Mini). This shift ensures full transparency and cost efficiency, replacing the old flat execution fee model with one that scales directly with usage.

IMPORTANT: Pricing has shifted from a fixed execution fee to a token-based, zero-markup model. Your costs now depend on:

1. The input tokens used in your prompts
2. The output tokens generated in the response
3. The AI model you choose

This means you pay OpenAI's exact rates with nothing more, nothing less.

External AI Models - Based on Tokens utilized

GPT-4o (per million tokens)
Input - $2.5
Output - $10.00
GPT-4o-mini (per million tokens)
Input - $0.15
Output - $0.60

For example,

With just $10 in usage, GPT-4o can process around 3 million input words or generate about 750,000 output words, while GPT-4o Mini offers far greater volume—handling roughly 50 million input words or producing about 12.5 million output words. This makes it easy to balance cost and performance, choosing the model that best matches your workflow’s needs.

What $10 Gets You

GPT-4o:

Process approximately 3 million input words OR
Generate approximately 750,000 output words

GPT-4o Mini:

Process approximately 50 million input words OR
Generate approximately 12.5 million output words

Use Cases and Example Prompts

Use Cases	Prompts
Lead Follow-Up	“Draft a friendly message to {{contact.name}} thanking them for their interest in {{contact.custom_value.service}}.”
Conversation Summary	“Summarize this conversation in 2-3 bullet points.”
Email Draft	“Write a persuasive email promoting our {{contact.custom_value.offer}} to a potential client.”

GPT Models Compared

GPT 3.5 Turbo	GPT-4 Turbo	GPT-4o	GPT-4o Mini
Not as good as 4 turbo	More advanced and powerful	Optimized for specific tasks	Lightweight version of GPT-4o
Good responses	Consistent and accurate responses	Consistent with slight trade-offs	Efficient but less comprehensive
Not that consistent	Can handle highly complex queries	Handles most queries effectively	Best for less complex queries
May struggle to connect the dots	A bit slower than GPT-4o Mini	Balanced speed and performance	Fastest in response time
Faster responses	Superior contextual understanding	Good contextual understanding	Adequate for basic needs
as per tokens utilized	as per tokens utilized	as per tokens utilized	$0.015 per execution

The GPT-Generated Output

After selecting and configuring the desired AI model, incorporate a 'GPT powered by OpenAI' action into your workflow to generate dynamic output aligned with your configured prompts."

Select the Action Type:
- Pick from predefined options:
  - Analyze Text Sentiment
  - Summarize Text
  - Translate
  - Custom (for your own prompt)
- If using Custom, enter your own prompt (e.g., "Write a follow-up email based on this message: {{contact.message}}").
Use Dynamic Variables (Optional):
- Insert dynamic fields like {{contact.first_name}} or {{event.notes}} into your prompt to personalize the output.
Test or Run the Workflow:
- Once the action is added, you can test the workflow using a contact to see the GPT output in action.
- The AI output can be used in subsequent workflow steps (e.g., send as an email or SMS).

IMPORTANT: If you are concerned about where does the output go,

The AI-generated content can be:-

1. Stored in Custom Fields

2. Used directly in Email or SMS steps

3. Logged internally or sent to a team via Slack/Webhook

Frequently Asked Questions

Q. What is GPT-4 Turbo, and how is it different from GPT-4 or GPT-3.5?
GPT-4 Turbo is a faster, more cost-efficient variant of GPT-4 with improved performance for high-volume use cases. Compared to GPT-3.5, it offers better contextual understanding, more accurate content generation, and supports larger prompts.

Q. Is GPT-4 Turbo available in all HighLevel plans?
No. GPT-4 Turbo is part of HighLevel’s premium Content AI features. Access may vary based on your subscription plan and usage settings. Check your plan or contact support for upgrade options.

Q. Does using GPT-4 Turbo in workflows cost extra?
Yes. GPT-4 Turbo is a premium workflow action, and usage is billed based on the number of tokens processed per execution. You can monitor usage and costs in the AI Billing Settings of your account.

Q. Can I switch between GPT-3.5 and GPT-4 Turbo within the same workflow?
Yes. Each AI step allows you to select the model independently. You can mix and match GPT-3.5, GPT-4, and GPT-4 Turbo within a single workflow based on the complexity or budget of each step.

Q. How many words are present in 1 million tokens?

This is an Estimation, actual results may vary: 1,000,000 tokens × 0.75 words/token = 750,000 words