Generative AI

ChatGPT: Some new features are now available for an additional fee

Since its launch in late 2022, ChatGPT has established itself as one of the most widely used tools in generative artificial intelligence. With over 100 million monthly active users, the model developed by OpenAI has rapidly become widespread in everyday, professional, and academic use. But a recent development is changing the game: certain new features will only be available for an additional fee.

This decision marks a strategic milestone for OpenAI, which is adapting its business model in response to the high costs of developing, hosting, and training its models. It also raises fundamental questions about accessibility, equity, and the future of AI tool distribution.

Until now, ChatGPT operated on a relatively straightforward model: a free version with limited capabilities and a paid “Plus” plan (around 20 euros per month) that provided access to the latest models. Going forward, OpenAI is introducing a more granular approach. Certain features will only be available for an additional fee, separate from the main subscription.

The idea is to distinguish between basic uses (conversation, writing, research assistance) and advanced features (intelligent agents, transactional capabilities, multimodality). This model is based on the principle of the scalable freemium model, which has already proven successful in collaborative software and cloud services: access is simple and often free, but specialized features require a fee.

According to a PwC report (2024), nearly 65% of technology companies that have integrated AI now use a modular pricing model to align operating costs with the diversity of use cases1. OpenAI is therefore no exception to this trend.

Among the new features that may incur additional costs, several have already been mentioned:

  • Smart agents: these extensions are capable of performing actions (booking a flight, making an online purchase, automating complex tasks).
  • The Instant Checkout transaction mode, which allows ChatGPT to complete a payment without leaving the chatbot.
  • Advanced multimodal capabilities: image analysis, generation of professional-quality visuals, transcription, and real-time speech synthesis.
  • Enhanced office integrations: such as “Agent Mode” in Word and Excel, which could be moved to a premium model.

OpenAI attributes this segmentation to the energy and hardware costs associated with such features. According to the University of Washington, training a large multimodal model can require several million dollars in computing power and consume as much electricity as a small town would use over several weeks2.

This tiered pricing has significant implications:

  • For individual users: this could create a barrier to accessing innovations by reserving certain features for premium subscribers.
  • For businesses: this follows a familiar model, in which business software is offered through subscription tiers.
  • For students and teachers: the issue of digital equity arises, particularly in educational settings where ChatGPT is increasingly being used as a learning tool.

A UNESCO report (2024) highlights that affordability remains a major challenge: 41% of the educational institutions surveyed consider the costs associated with generative AI to be a barrier to its integration into teaching3.

With this decision, OpenAI is pursuing a twofold goal:

  • Getting a return on its massive investments. In 2024, the company raised over $1 billion to fund its infrastructure and accelerate research.
  • Diversify revenue streams. The ChatGPT Plus subscription is not enough to cover the costs associated with cloud infrastructure, which are estimated at several hundred million dollars a year.

This strategy brings OpenAI closer to the business models of major digital platforms: offering a widely available basic service to create a network effect, and then restricting access to the most expensive innovations.

The introduction of additional fees also raises questions of transparency and fairness.

  • Pricing transparency: Users must clearly understand which features remain included and which ones now require payment.
  • Equitable access: a digital divide is emerging between those who can afford advanced features and those who will be limited to a stripped-down version.
  • Ethics: Should AI used in education, the workplace, or research remain accessible without significant additional costs?

According to a BEUC study (2024), 57% of European consumers believe that the additional fees charged by tech companies for access to AI could widen inequalities in access to digital tools4.

With the introduction of additional fees, OpenAI confirms that the future of artificial intelligence will be shaped by à la carte pricing. Widespread access to basic features will coexist with advanced options reserved for those willing to pay more.

This strategy reflects a broader trend: AI is evolving into a multi-tiered market, where the most advanced innovations are becoming premium services. For users, the challenge will be to strike a balance between the tangible benefits and the cost of these services. For institutions, it poses a challenge in terms of regulation and democratization.

OpenAI is betting on the rise of AI integrated into professional and personal practices, though access to features beyond the basics will depend on individual ability to pay. It remains to be seen whether this model will gain social acceptance, and whether artificial intelligence will become a public good or a market-driven product in the future.

In a previous post on this blog, we analyzed the evolution of the AI business model with: OpenAI is betting big: $1.1 billion to take its AI to the next level

1. PwC. (2024). AI Monetization Strategies in Tech.
https://www.pwc.com

2. University of Washington. (2023). Energy Costs of Large AI Models.
https://www.washington.edu

3. UNESCO. (2024). AI in Education: Access and Equity Report.
https://www.unesco.org

4. BEUC. (2024). Consumers and Artificial Intelligence: Pricing and Fair Access.
https://www.beuc.eu

Don't miss our upcoming articles!

Get the latest articles written by aivancity experts and professors delivered straight to your inbox.

We don't send spam! Please see our privacy policy for more information.

Don't miss our upcoming articles!

Get the latest articles written by aivancity experts and professors delivered straight to your inbox.

We don't send spam! Please see our privacy policy for more information.

Related posts
Generative AI

OpenAI unveils GPT-5.4, a model designed for complex reasoning and coding

GPT-5.4 is available in two main versions: GPT-5.4 Thinking and GPT-5.4 Pro. Both versions are based on the same architecture but differ in terms of performance, speed, and pricing. One of the advancements…
Generative AI

Nano Banana 2: Google Accelerates Image AI at Lightning Speed

Google is continuing its push into generative visual AI with the launch of Nano Banana 2, also known as Gemini 3.1 Flash Image. This new model does more than just improve…
Generative AI

Gemini 3.1 Pro: Google's answer to the most advanced models on the market

Google is continuing to ramp up its strategic push into generative artificial intelligence with the launch of Gemini 3.1 Pro, a version touted as significantly more powerful than its predecessor. Against a backdrop of intense competition among the major players…
The AI Clinic

Would you like to submit a project to the AI Clinic and work with our students?

Leave a comment

Your email address will not be published. Required fields are marked with *