GPT Store

See also: Custom GPTs, ChatGPT, and OpenAI

The GPT Store is OpenAI's marketplace for discovering, sharing, and distributing custom versions of ChatGPT known as Custom GPTs.^[1] It launched on January 10, 2024, and by that date users had already created over 3 million Custom GPTs, of which roughly 159,000 were published publicly in the store.^[3][7] Announced on November 6, 2023, at OpenAI's inaugural DevDay conference in San Francisco, the GPT Store allows anyone with a ChatGPT Plus, Team, or Enterprise subscription to browse, use, and publish specialized AI assistants built on top of OpenAI's GPT-4 and later models.^[1][3] In announcing the launch, OpenAI wrote: "It's been two months since we announced GPTs, and users have already created over 3 million custom versions of ChatGPT."^[3]

Custom GPTs require no programming. OpenAI describes them as "a new way for anyone to create a tailored version of ChatGPT to be more helpful in their daily life, at specific tasks, at work, or at home, and then share that creation with others."^[1] Adoption has since shifted toward the workplace: in OpenAI's "The State of Enterprise AI" report published December 8, 2025, weekly users of Custom GPTs and Projects had grown roughly 19x year to date, and approximately 20% of all Enterprise messages were processed through a Custom GPT or Project.^[21]

History

What was announced at DevDay (November 6, 2023)?

OpenAI CEO Sam Altman introduced the concept of Custom GPTs during the keynote at OpenAI's first-ever DevDay developer conference, held on November 6, 2023, in San Francisco.^[1] Alongside the announcement of GPT-4 Turbo and a new Assistants API, Altman demonstrated how users could create tailored versions of ChatGPT for specific purposes without writing any code.^[2] The demonstration showed the GPT Builder interface (originally codenamed "Magic Maker" during internal development) in action, walking attendees through the process of building a custom GPT in real time through natural language conversation.^[1]

At the time of the DevDay announcement, OpenAI stated that a GPT Store would launch "later this month."^[1] However, the store's release was delayed due to internal leadership turmoil at OpenAI in November 2023, when the board briefly removed Sam Altman as CEO before reinstating him days later. The GPT Store launch was pushed back to January 2024.^[5]

When did the GPT Store launch?

The GPT Store officially went live on January 10, 2024.^[3][4] At launch, OpenAI reported that users had already created over 3 million custom GPTs in the roughly two months since the feature was announced.^[3] The store was initially available to ChatGPT Plus ($20/month), Team ($25/user/month), and Enterprise subscribers.^[4]

Alongside the GPT Store, OpenAI also introduced the ChatGPT Team plan, a new subscription tier priced at $25 per user per month (billed annually) or $30 per user per month (billed monthly), designed for small and medium-sized teams that wanted collaborative workspace features and higher usage limits.^[5][13]

The store featured curated selections from OpenAI's editorial team, community leaderboards highlighting trending GPTs, and category-based browsing to help users find relevant custom assistants.^[3]

Expansion to Free Users (May 2024)

On May 13, 2024, OpenAI announced the launch of GPT-4o and expanded access to a range of tools for free-tier ChatGPT users.^[6] As part of this rollout, free users gained the ability to browse and use GPTs from the GPT Store, though they could not create or edit GPTs without a paid subscription.^[6] Free users also faced stricter rate limits on advanced model usage.^[6]

How Custom GPTs Work

Overview

Custom GPTs are specialized versions of ChatGPT that combine a custom system prompt, optional uploaded knowledge files, and configurable capabilities (such as web browsing, DALL-E image generation, and Code Interpreter) into a single, shareable package.^[1][8] They run on OpenAI's GPT models and are accessed through the ChatGPT interface.

When a user interacts with a Custom GPT, the system assembles the conversation context by combining the creator's instructions, any relevant passages retrieved from uploaded knowledge files, and the user's messages. The model then generates responses guided by all of these inputs.

GPT Builder

The GPT Builder is OpenAI's no-code creation tool for building Custom GPTs.^[8] Rather than requiring users to write code or configure technical settings manually, the GPT Builder operates as a conversational interface. Users describe what they want their GPT to do in plain language, and the builder helps them refine the instructions, suggest a name and profile picture, and configure capabilities.^[8]

The creation process involves the following steps:

1. Start a conversation with the GPT Builder, describing the desired behavior and purpose.
2. Refine instructions through follow-up dialogue, specifying tone, expertise domains, and behavioral rules.
3. Upload knowledge files (optional) to give the GPT access to specific reference material.
4. Configure capabilities such as web browsing, image generation with DALL-E, and code execution.
5. Set up Actions (optional) to connect the GPT to external APIs.
6. Test and publish the GPT, choosing between private, link-shared, or public visibility.

For users who prefer more direct control, the GPT Builder also provides a Configure tab where all settings (instructions, knowledge files, capabilities, and actions) can be edited manually.^[8]

Technical Architecture

Under the hood, Custom GPTs use a model variant sometimes referred to internally as gpt-4-gizmo, running within a 32,000-token (and later 128,000-token) context window shared across the conversation. The technical components include:

Scroll sideways for more →

Knowledge Retrieval (RAG)

When a creator uploads files to a Custom GPT, OpenAI's system indexes the content and uses retrieval-augmented generation (RAG) to find relevant passages at query time. The GPT does not memorize the entire contents of uploaded files. Instead, when a user asks a question, the system performs a semantic search over the indexed documents, retrieves the most relevant chunks, and injects them into the model's context window alongside the conversation history and system instructions.

Supported file types for knowledge uploads include PDFs, Word documents, text files, spreadsheets, presentations, and code files.^[8] This approach mirrors traditional RAG pipelines used in enterprise AI applications, where a vector store and retriever fetch semantically matching content to ground the model's responses in specific source material.

GPT Actions

GPT Actions allow Custom GPTs to interact with external services and APIs using natural language. They build on OpenAI's function calling capability to translate a user's conversational request into a structured API call.^[9]

The process works as follows:

1. The GPT creator defines an API endpoint using an OpenAPI specification (a standardized format for describing REST APIs).
2. The creator provides a plain-language description of what the API does and when the GPT should use it.
3. During a conversation, the model evaluates whether the user's request requires an API call.
4. If so, the model generates the appropriate JSON payload, executes the API call, and incorporates the response into its reply.^[9]

Actions support multiple authentication mechanisms, including API keys, OAuth 2.0, and no-auth endpoints.^[9] Common use cases include querying databases, filing support tickets, retrieving real-time data from external services, and triggering workflows in tools like Zapier, Slack, or Jira.

OpenAI provides a GPT Actions library with pre-built templates for popular integrations, making it easier for creators to connect their GPTs to third-party services without deep API expertise.^[9]

GPT Store Categories and Discovery

The GPT Store organizes Custom GPTs into several main categories to help users discover relevant tools:

Scroll sideways for more →

Discovery within the store relies on a combination of editorial curation, community rankings based on conversation volume, and a search function. OpenAI features trending GPTs on the store's homepage and maintains leaderboards that surface the most popular GPTs in each category.^[3]

Popular Custom GPTs

Some of the most widely used GPTs in the store include:

• Consensus: A research assistant that searches through and synthesizes insights from published academic papers.
• Write For Me: A writing GPT specialized in generating tailored, engaging content with precise word count control.
• Python: A highly specialized GPT optimized for Python programming, code execution, and data analysis.
• Canva: An integration with the Canva design platform for creating visual content.
• Code Copilot: A programming assistant offering code generation, debugging, and explanation.
• Scholar GPT: A GPT focused on academic research with access to scholarly databases.
• Wolfram: Access to Wolfram Alpha's computational knowledge engine and Wolfram Language.
• DesignerGPT: Creates and hosts websites, integrating DALL-E-generated images with web development tools.

Revenue Sharing Program

When the GPT Store launched in January 2024, OpenAI announced plans for a GPT Builder Revenue Program to compensate creators based on user engagement with their GPTs.^[3][10] The company stated it would launch the program in Q1 2024, starting with US-based builders.^[10]

Eligibility Requirements

To qualify for the revenue sharing program, GPT creators must meet the following criteria:

Scroll sideways for more →

How much do GPT creators actually earn?

The revenue sharing program did not launch broadly as initially promised. Instead of a full public rollout in Q1 2024, OpenAI implemented a limited, invite-only pilot program for select builders.^[10][20] Developer Nick Dobos, creator of the popular Grimoire GPT, was reportedly invited in March 2024 and received a guaranteed minimum payout of $1,000 per month, plus additional earnings based on usage.^[19]

Based on community reports, the earnings landscape for most creators has been modest:^[20]

Scroll sideways for more →

The program's limited scope and lack of transparency have drawn criticism from the creator community.^[20] Many builders have expressed frustration on OpenAI's developer forums about the delayed rollout and unclear payment criteria. As of early 2026, OpenAI has not expanded the program into a fully public, broadly accessible revenue sharing system.

GPTs vs. ChatGPT Plugins

Background

ChatGPT Plugins were introduced by OpenAI in March 2023 as an early mechanism for extending ChatGPT's capabilities through third-party integrations.^[11] Plugins allowed ChatGPT to access external tools, services, and data sources during conversations. Over 1,000 plugins were made available, including integrations with services like Expedia, Instacart, Kayak, and Wolfram Alpha.^[11]

Deprecation of Plugins

OpenAI deprecated ChatGPT Plugins in early 2024, replacing them with Custom GPTs and GPT Actions.^[11] The deprecation followed this timeline:

Scroll sideways for more →

Reasons for the Transition

Several factors drove OpenAI's decision to replace plugins with Custom GPTs:

• Low adoption: Despite having over 1,000 plugins, usage remained concentrated among a small group of power users. Most ChatGPT Plus subscribers never explored or understood how plugins worked.^[11]
• Fragmented experience: Installing and managing multiple plugins created a confusing user experience, especially for less technical users.
• Native capabilities: Many plugin functions (web browsing, code execution, image generation, file analysis) were being built directly into ChatGPT, reducing the need for third-party plugins.
• Superior flexibility: Custom GPTs offered a more unified and customizable approach, combining instructions, knowledge, and API integrations into a single package rather than relying on separate plugin installations.

Key Differences

Scroll sideways for more →

Enterprise and Team Custom GPTs

Enterprise Features

OpenAI offers enhanced GPT management and governance features for organizations using ChatGPT Enterprise and ChatGPT Edu (for educational institutions).^[12] These features give workspace administrators granular control over how GPTs are created, shared, and used within their organizations.^[12]

Scroll sideways for more →

How widely are Custom GPTs used in the enterprise?

Custom GPTs have become a significant share of business AI usage. In OpenAI's "The State of Enterprise AI" report, published December 8, 2025 and based on a survey of 9,000 workers across nearly 100 enterprises, structured workflows such as Projects and Custom GPTs increased roughly 19x year to date, and approximately 20% of all Enterprise messages were processed through a Custom GPT or Project.^[21] OpenAI framed this as evidence that teams are "packaging repeatable work into shareable assets" and then scaling that behavior across functions.^[21]

ChatGPT Team (now Business) Plan

The ChatGPT Team plan, launched alongside the GPT Store in January 2024, was designed specifically for smaller teams that needed collaborative GPT features without the full Enterprise contract.^[13] Priced at $25 per user per month (billed annually), it included:^[13]

• Access to GPT-4 and DALL-E 3 with higher message limits
• A shared workspace for team-created GPTs
• Data excluded from model training
• Admin tools for user management and billing
• Minimum of two seats required

In mid-2025, OpenAI rebranded the Team plan as ChatGPT Business, retaining the same core features and pricing structure.

Comparison with Claude Projects

Anthropic's Claude offers a feature called Claude Projects that serves a similar purpose to Custom GPTs: allowing users to create customized AI experiences with specific instructions and knowledge. However, the two approaches differ in several important ways.^[14]

Scroll sideways for more →

Claude Projects excel in handling large documents and long-context tasks thanks to Anthropic's larger context windows, while Custom GPTs are better suited for creating distributable, public-facing AI tools with API integrations and a built-in marketplace for discovery.^[14]

Statistics

Scroll sideways for more →

Evolution: 2025 and Beyond

Apps in ChatGPT and the Apps SDK (October 2025)

On October 6, 2025, OpenAI introduced a new concept called Apps in ChatGPT along with an accompanying Apps SDK (software development kit) in developer preview.^[15] While Custom GPTs are prompt-based tools that package behavior and instructions, Apps represent a deeper level of integration. Apps can include backend logic, external API calls, and interactive widgets, going beyond what GPTs can do through Actions alone.^[15]

The key distinction between GPTs and Apps:

Scroll sideways for more →

ChatGPT App Directory (December 2025)

On December 18, 2025, OpenAI launched a dedicated App Directory inside ChatGPT, available to Plus, Team (Business), and Enterprise users.^[16][17] The directory allows users to browse approved apps and add them directly to their workspace. Developers can submit apps for review through a new developer portal, with approved apps appearing in the in-product directory for discovery.^[17]

Initial launch partners included integrations from Apple Music (music curation), DoorDash (food ordering), Zillow (real estate search), and Khan Academy (interactive learning), demonstrating the range of experiences possible within the directory.^[16]

Apps that pass OpenAI's review process began rolling out to users in early 2026, signaling a shift in how OpenAI envisions the ecosystem around ChatGPT.^[16] The GPT Store continues to operate alongside the App Directory, with GPTs serving as lightweight, no-code customizations and Apps offering more robust, developer-built integrations.

Looking Ahead

The GPT Store and the broader Custom GPTs ecosystem represent OpenAI's effort to build a platform layer on top of ChatGPT, similar to how Apple's App Store transformed the iPhone from a single product into a platform. However, the execution has faced challenges, including the delayed and limited revenue sharing program, quality control issues with the volume of GPTs submitted, and competition from alternative approaches like Anthropic's Claude Projects and Google's Gemini Gems.^[18]

As OpenAI continues to develop the App Directory and Apps SDK alongside the existing GPT Store, the line between Custom GPTs and full-featured apps will likely continue to blur, with both serving different segments of the creator and developer community.

Popular GPTs Table

The table below lists notable Custom GPTs, ordered by their popularity (number of chats). The most popular GPTs are listed first.

Scroll sideways for more →

See also

• EditGPT
• BrowsingGPT
• ExcelGPT
• GPT Customizer, File Finder & JSON Action Creator
• AI Humanizer by BypassGPT

References

1. OpenAI. "Introducing GPTs." OpenAI Blog, November 6, 2023. https://openai.com/index/introducing-gpts/ ↩
2. OpenAI. "New models and developer products announced at DevDay." OpenAI Blog, November 6, 2023. https://openai.com/index/new-models-and-developer-products-announced-at-devday/ ↩
3. OpenAI. "Introducing the GPT Store." OpenAI Blog, January 10, 2024. https://openai.com/index/introducing-the-gpt-store/ ↩
4. Wiggers, Kyle. "OpenAI launches a store for custom AI-powered chatbots." TechCrunch, January 10, 2024. https://techcrunch.com/2024/01/10/openai-launches-a-store-for-custom-ai-powered-chatbots/ ↩
5. Novet, Jordan. "OpenAI debuts GPT Store, new business tier after leadership fracas." CNBC, January 10, 2024. https://www.cnbc.com/2024/01/10/openai-debuts-gpt-store-new-business-tier-after-leadership-fracas.html ↩
6. OpenAI. "Introducing GPT-4o and more tools to ChatGPT free users." OpenAI Blog, May 13, 2024. https://openai.com/index/gpt-4o-and-more-tools-to-chatgpt-free/ ↩
7. SEO.ai. "GPT Store Statistics & Facts." SEO.ai Blog. https://seo.ai/blog/gpt-store-statistics-facts ↩
8. OpenAI. "Creating a GPT." OpenAI Help Center. https://help.openai.com/en/articles/8554397-creating-a-gpt ↩
9. OpenAI. "GPT Actions." OpenAI Developer Documentation. https://developers.openai.com/api/docs/actions/introduction ↩
10. VentureBeat. "OpenAI launches GPT Store but revenue sharing is still to come." VentureBeat, January 10, 2024. https://venturebeat.com/ai/openai-launches-gpt-store-but-revenue-sharing-is-still-to-come ↩
11. Practical-Tech. "ChatGPT Plugins are being replaced by GPTs." March 7, 2024. https://practical-tech.com/2024/03/07/chatgpt-plugins-are-being-replaced-by-gpts-heres-why-and-what-it-means-for-you/ ↩
12. OpenAI. "Managing GPT access in Enterprise and Edu workspaces." OpenAI Help Center. https://help.openai.com/en/articles/8555535-gpts-chatgpt-enterprise-version ↩
13. OpenAI. "Introducing ChatGPT Team." OpenAI Blog. https://openai.com/index/introducing-chatgpt-team/ ↩
14. Bowdoin, Jeffrey. "Claude Projects vs. Custom GPTs: A Comprehensive Comparison." https://jeffreybowdoin.com/claude-projects-vs-custom-gpts/ ↩
15. OpenAI. "Introducing apps in ChatGPT and the new Apps SDK." OpenAI Blog, October 6, 2025. https://openai.com/index/introducing-apps-in-chatgpt/ ↩
16. Lardinois, Frederic. "ChatGPT launches an app store, lets developers know it's open for business." TechCrunch, December 18, 2025. https://techcrunch.com/2025/12/18/chatgpt-launches-an-app-store-lets-developers-know-its-open-for-business/ ↩
17. OpenAI. "Developers can now submit apps to ChatGPT." OpenAI Blog. https://openai.com/index/developers-can-now-submit-apps-to-chatgpt/ ↩
18. Originality.AI. "The Era of Tailored Intelligence: Charting the Growth and Market Impact of Custom GPTs." https://originality.ai/blog/gpts-statistics ↩
19. Dobos, Nick (@NickADobos). "More details on openAi revenue share for GPT Store: -$1,000 USD/mo minimum -potential for additional earnings based on usage." X, March 28, 2024. https://x.com/NickADobos/status/1773456722675535957 ↩
20. The GPT Shop. "OpenAI GPT Store Revenue Sharing Explained: What Creators Actually Earn." The GPT Shop Blog. https://www.thegptshop.online/blog/openai-gpt-store-revenue-sharing ↩
21. OpenAI. "The state of enterprise AI." OpenAI, December 8, 2025. https://openai.com/index/the-state-of-enterprise-ai-2025-report/ ↩