Introducing study mode Introducing study mode

Try in ChatGPT(opens in a new window)

Today we’re introducing study mode in ChatGPT—a learning experience that helps you work through problems step by step instead of just getting an answer. Starting today, it’s available to logged in users on Free, Plus, Pro, Team, with availability in ChatGPT Edu coming in the next few weeks.

ChatGPT is becoming one of the most widely used learning tools in the world. Students turn to it to work through challenging homework problems, prepare for exams, and explore new concepts. But its use in education has also raised an important question: how do we ensure it is used to support real learning, and doesn’t just offer solutions without helping students make sense of them?

We’ve built study mode to help answer this question. When students engage with study mode, they’re met with guiding questions that calibrate responses to their objective and skill level to help them build deeper understanding. Study mode is designed to be engaging and interactive, and to help students learn something—not just finish something.

How study mode was built

Under the hood, study mode is powered by custom system instructions we’ve written in collaboration with teachers, scientists, and pedagogy experts to reflect a core set of behaviors that support deeper learning including: ​​encouraging active participation, managing cognitive load, proactively developing metacognition and self reflection, fostering curiosity, and providing actionable and supportive feedback. These behaviors are based on longstanding research in learning science and shape how study mode responds to students.

_“Instead of doing the work for them, study mode encourages students to think critically about their learning. Features like these are a positive step toward effective AI use for learning. Even in the AI era, the best learning still happens when students are excited about and actively engaging with the lesson material.”—Robbie Torney, Senior Director of AI Programs at Common Sense Media._

Key features

Interactive prompts:Combines Socratic questioning, hints, and self-reflection prompts to guide understanding and promote active learning, instead of providing answers outright.

Scaffolded responses:Information is organized into easy-to-follow sections that highlight the key connections between topics, keeping information engaging with just the right amount of context and reducing overwhelm for complex topics.

Personalized support:Lessons are tailored to the right level for the user, based on questions that assess skill level and memory from previous chats.

Knowledge checks:Quizzes and open-ended questions, along with personalized feedback to track progress, support knowledge retention and the ability to apply that knowledge in new contexts.

Flexibility:Easily toggle study mode on and off during a conversation, giving you the flexibility to adapt to your learning goals in each conversation.

How students use study mode

Study mode was built with college students in mind. Some feedback we heard from students during early testing:

_“The best way I’d describe it is a live, 24/7, all-knowing “office hours.” —Noah Campbell, college student_

“_Study mode did a great job breaking down dense material into clear, well-paced explanations.” —Caleb Masi, college student_

_“I put study mode to the test to tutor me on a concept I have attempted to learn many times before: sinusoidal positional encodings. It was like a tutor who doesn’t get tired of my questions. After a 3-hour working session, I finally understood it well enough to feel confident.”—Maggie Wang, college student_

Ways to use study mode inspired by college students from the ChatGPT lab

Parker Jones (Game Theory)Praja Tickoo (Accounting)Nolan Windham (Bayes' Theorem)Cecily Chung (Discrete Mathematics)

I want to learn about Game Theory, specifically the broad spectrum that the field entails, then also the ways in which you think it'd be useful for me to understand in my daily life. I want you to of course follow my curiousity, but mostly you will be teaching me about it, and keeping a high level plan to iterate through so I can cover the full scope here. I will ask questions when I am curious, but be deadset on quickly educating me on this.

Perfect. Here’s the high-level plan I’ll drive us through—think of it as a layered roadmap from foundational ideas to practical intuition:

Perfect. Here’s the high-level plan I’ll drive us through—think of it as a layered roadmap from foundational ideas to practical intuition:

🔹 Phase 1: Core Foundations

1. What is Game Theory, really?

1. What is Game Theory, really?

2. Types of Games

2. Types of Games

* Cooperative vs non-cooperative

* Simultaneous vs sequential

3. Basic Concepts

3. Basic Concepts

* Players, strategies, payoffs, information

* Dominant strategies

🔸 Phase 2: Classic Games & Concepts

4. Famous Games

🔸 Phase 2: Classic Games & Concepts

4. Famous Games

* Prisoner’s Dilemma

* Chicken

* Battle of the Sexes

5. Mixed Strategy Equilibria

* Matching Pennies

5. Mixed Strategy Equilibria

🔹 Phase 3: Deepening the Model

6. Repeated Games & Folk Theorems

* * *

🔹 Phase 3: Deepening the Model

7. Extensive-Form Games

* Cooperation emerging over time

* Tit for Tat, Grim Trigger, etc.

8. Bayesian Games

* Game trees, backward induction

* Subgame Perfect Equilibria

🔸 Phase 4: Real-Life Relevance

9. Mechanism Design / Reverse Game Theory

* Beliefs, types, and expected utility

10. Behavioral Game Theory

🔸 Phase 4: Real-Life Relevance

11. Daily Life Applications

* Voting systems, auctions, contracts

🔹 Phase 5: Advanced & Optional

12. Evolutionary Game Theory

11. Daily Life Applications

13. Algorithmic Game Theory

* * *

14. Your world: How it matters to Parker

12. Evolutionary Game Theory

Throughout this, I’ll mix in:

13. Algorithmic Game Theory

* Network effects, internet auctions

14. Your world: How it matters to Parker

* Systems design, incentives, power structures

Let’s kick off Phase 1.

First quick one: In your own words, what do you think Game Theory is about? No need to be polished—just get it out.

I think Game Theory is sort of this meta aspect framing of modern society in life where we look at people in situations and complex systems and we try to analyze how people in them make decisions, sometimes why they make it, and how this affects the situation/system

Nice—you're already circling the core.

Let’s sharpen it a bit together. Game Theory is fundamentally about strategic decision-making—situations where the outcome for each person depends not just on what _they_ do, but on what _others_ do too.

So real quick:

> If you’re making a decision and it depends entirely on nature or chance—like a dice roll—is that Game Theory?

And then:

> If you’re making a decision and it depends on what someone else does—like whether they cooperate or betray you—is that Game Theory?

Walk me through your thinking on both.

No random chance isn't game theory, though agents in game theory may end up having things depend on random chance. But if youre decision matters then of course it is. Also please remember you should be teaching this to me, via paragraphs of information and moving through a high level plan. If I want to do interactions and questions, I will ask. Continue

Got it—I'll stay in teaching mode and drive forward through the roadmap unless you tap in.

📍Phase 1: Core Foundations — What Game Theory_Is_

At its heart, Game Theory studies strategic interactions: any situation where the outcome for each agent (person, company, nation, etc.) depends on the actions of others. It’s not about games in the traditional sense, but rather any “game” where decisions must be made in the presence of others who are also making decisions.

Key Assumptions (Classical Game Theory):

1. Rational agents – Everyone acts to maximize their own utility (whatever that means to them).

2. Strategic thinking – Agents consider what others might do before acting.

3. Common knowledge – Everyone knows the rules, the payoffs, and that everyone is rational (though later variants relax this).

🧩 Core Building Blocks

Here are the elements every formal game model needs:

📍Phase 1: Core Foundations — What Game Theory_Is_

At its heart, Game Theory studies strategic interactions: any situation where the outcome for each agent (person, company, nation, etc.) depends on the actions of others. It’s not about games in the traditional sense, but rather any “game” where decisions must be made in the presence of others who are also making decisions.

Key Assumptions (Classical Game Theory):

1. Rational agents – Everyone acts to maximize their own utility (whatever that means to them).

These get framed into a “game form”, and from there, you look for solutions—most famously, Nash equilibrium (more on that soon).

🎲 Quick Clarification: Game Theory vs Probability

You're exactly right—pure chance (like dice) isn’t Game Theory. That’s probability theory. But if _you_ are choosing a strategy knowing that _others_ may randomize their own moves? That’s Game Theory.

For example: in rock-paper-scissors, the “optimal” strategy is _not deterministic_—it’s a mixed strategy, which involves randomizing your choices. But you’re doing so _strategically_, in response to an opponent.

🧠 Where We’re Going Next

Next up, I’ll walk through types of games (Phase 1 continued), then move into dominant strategies and Nash equilibrium to set up the first actual game analysis.

Let me know if you want to pause, go deeper, or shift gears—but otherwise I’ll keep pressing forward.

I like the rock paper scissors example. Keep going, connecting to my life where possible. want to internalize this to be a better person/decision maker

Getting started

Study mode is especially useful for homework help, test prep, and learning new topics. To try it, select “Study and learn” from tools in ChatGPT and ask a question.

Working through a homework problem Prepping for an exam

Limitations and what’s next

This is a first step in a longer journey to improve learning in ChatGPT. Today, study mode is powered by custom system instructions. We chose this approach because it lets us quickly learn from real student feedback and improve the experience—even if it results in some inconsistent behavior and mistakes across conversations. We plan on training this behavior directly into our main models once we’ve learned what works best through iteration and student feedback.

We’re also exploring functionality to make study mode more engaging and helpful for students including:

* * *

🧠 Where We’re Going Next

Next up, I’ll walk through types of games (Phase 1 continued), then move into dominant strategies and Nash equilibrium to set up the first actual game analysis.

The research landscape is still taking shape on the best ways to apply AI in education. OpenAI is enabling further research on learning and AI through various partners in its NextGenAI⁠ initiative, and working with experts from the SCALE Initiative at Stanford University’s Accelerator for Learning⁠(opens in a new window) to study and share how AI tools influence learning outcomes in areas like K-12 education.

As we run longer-term studies on how students learn best with AI, we intend to publish a deeper analysis of what we’ve learned about the links between model design and cognition, shape future product experiences based on these insights, and work side by side with the broader education ecosystem to ensure AI benefits learners worldwide.

Getting started

Study mode is especially useful for homework help, test prep, and learning new topics. To try it, select “Study and learn” from tools in ChatGPT and ask a question.

Working through a homework problem Prepping for an exam

Author

OpenAI

We’re also exploring functionality to make study mode more engaging and helpful for students including:

* Clearer visualizations for complex or text-heavy concepts

* Goal setting and progress tracking across conversations

* Deeper personalization tailored to each student’s skill level and goals

The research landscape is still taking shape on the best ways to apply AI in education. OpenAI is enabling further research on learning and AI through various partners in its NextGenAI⁠ initiative, and working with experts from the SCALE Initiative at Stanford University’s Accelerator for Learning⁠(opens in a new window) to study and share how AI tools influence learning outcomes in areas like K-12 education.

As we run longer-term studies on how students learn best with AI, we intend to publish a deeper analysis of what we’ve learned about the links between model design and cognition, shape future product experiences based on these insights, and work side by side with the broader education ecosystem to ensure AI benefits learners worldwide.

* 2025

* ChatGPT

* User Safety & Control

Author

OpenAI