Be there as OpenAI unveils what’s next. DevDay opens with a keynote packed with announcements—and new possibilities for what you can build with AI.
8:00 AM
BreakfastWalk In The Park
Getting inEnter the campus at Marina Boulevard and Buchanan Street and follow OpenAI DevDay signage.
10:00 AM
11:15 AM
What engineering work deserves your time and tokens? Learn from a Codex run that spent more than 1,000 hours keeping builds and tests passing in a large shared codebase. See how this internal OpenAI agent had goals, subagents, and persistent plans that helped it investigate failures and coordinate fixes. Leave with a playbook for choosing useful recurring tasks, improving test feedback, setting token budgets, and deciding when engineers should step in.
As models become more capable, agent architectures are changing. We’ll trace the evolution from simple loops to collaborating agents, examine the design choices and model advancements behind that shift, and take a deep dive into what comes next.
Stack Overflow’s VP of Product Alex Lato will share insights and practical learnings from bootstrapping and building with Codex. She’ll cover how two members of the Stack technical team created Stack Overflow for Agents, their API-first knowledge exchange, and how her team used the same principles to rapidly reimagine organizational knowledge with their enterprise product Stack Internal. Leave with lessons on using AI to accelerate R&D, modernize existing technology, and bring ambitious new products to market with small teams.
11:30 AM
LunchWalk In The Park
12:00 PM
Join Peter Steinberger for a look at how agents can change the way open-source teams build. Explore his team’s workflow for running agents on a shared server, giving them context across projects, and bringing external contributors into isolated environments. Through live demos, see how cloud sessions and computer use expand what agents can do. Leave with practical ideas for helping a small team build and collaborate at a larger scale.
Learn what it takes to run the Codex harness as a service through the Agents API. We’ll explore how agents work in your environment, how applications follow their activity, and how long-running tasks keep going. You’ll leave knowing which infrastructure OpenAI manages and where to focus your own engineering work.
ChatGPT enables people and teams to take on work that once felt out of reach. Step inside workflows from OpenAI’s power users. Follow a launch as it unfolds, explore competitive research, and see how our Chief People Officer turns expertise into tools. Leave knowing how to move faster, think bigger, and do more of your best work with ChatGPT.
12:45 PM
With Astra, ChatGPT Sites, and WebMCP, builders can now build new kinds of interfaces. Learn how to turn an idea into an interactive experience, and how you can collaborate with colleagues and agents to build the next generation of web apps.
Get more successful work from your AI budget. Drawing on customer deployments, this panel explores how to measure cost per completed task, choose the right model and reasoning effort, and reduce spend with prompt caching, Programmatic Tool Calling, and the Batch API. Learn how to spot wasted tokens, account for retries and human rework, and use evals to check that savings preserve quality. Leave with a practical approach to improving the cost of your own application.
1:00 PM
How did games shape AI? What makes a game fun? From Alan Turing to Astra, we’ll explore how games shaped modern AI. Then we’ll learn to build and playtest a retro-style 2D game with Codex, designing its characters, controls, and levels. You’ll see how AI can bring your ideas to life, and how your own taste and creativity make a game worth playing.
1:30 PM
Learn to use OpenAI’s latest cyber models and open-source CLI and SDK to find and fix vulnerabilities. Drawing on their experience securing OpenAI’s own codebase, Ian Webster and Kyle Brown will show how to scan large codebases, catch vulnerabilities in pull requests, and verify fixes.
Higgsfield CEO and co-founder Alex Mashrabov explains how an agent powered by OpenAI models chooses the right model for each task and provisions compute to run it. Learn when to allocate dedicated GPUs versus call a hosted model, how to deliver predictable throughput for enterprise workloads, and what fails when agents spend real compute on outputs users judge in seconds.
2:15 PM
Make ChatGPT the platform for what you’re building, whether it’s a personal tool, a team workflow, or a product for your users. Learn how plugin extensions bring your tools, knowledge, and interfaces natively into ChatGPT and Codex, and make them easily discoverable. Through live demos and a look at how OpenAI is building, see how to turn an idea into a useful experience, share it with your team, and reach new users.
As we move from delegating individual tasks to letting Codex tackle entire problems and goals on its own, it’s possible to harness the power of Codex wherever and however you work. Join us for a session packed with live demos covering what’s new in Codex in the CLI, app, and on the web, and find your new favorite way to build.
Tejal Patwardhan from the agents post-training team will share how OpenAI researchers are making models more useful and aligned, from computer use to agentic coding to complex knowledge work.
3:00 PM
Follow a 2am outage from the first Slack messages to a reviewed fix and an update for the team. Jason shows how voice, Appshots, plugins, and meeting notes help your persistent personal agent work through the investigation. Includes a preview of one of the DevDay launches.
See GPT-Live in action that can brainstorm, play games, and create images while coordinating work in the background. Then go under the hood to learn how to add voice to your own agents, manage delegation, and design for latency, interruptions, and changing requests.
Qodo CEO Itamar Friedman shares how his team builds with OpenAI to review AI-generated code. Explore the harness behind an adversarial reviewer designed to challenge a coding agent’s assumptions, reasoning, and output—from gathering organizational context to adversarial checks and validation workflows. Learn from the architectures, evals, tradeoffs, and failure modes shaping the system, including how to recognize when model confidence needs human judgment.
4:00 PM
Close out DevDay with Sam Altman, Tejal Patwardhan, and Thibault Sottiaux as they take questions from the community, share their perspectives on the day’s announcements, and discuss what’s next for AI and the developers building with it.
4:45 PM
ReceptionWalk In The Park
8:00 AM
BreakfastWalk In The Park
Getting inEnter the campus at Marina Boulevard and Buchanan Street and follow OpenAI DevDay signage.
10:00 AM
Festival Pavillion
Be there as OpenAI unveils what’s next. DevDay opens with a keynote packed with announcements—and new possibilities for what you can build with AI.
11:15 AM
Cowell Theater
What engineering work deserves your time and tokens? Learn from a Codex run that spent more than 1,000 hours keeping builds and tests passing in a large shared codebase. See how this internal OpenAI agent had goals, subagents, and persistent plans that helped it investigate failures and coordinate fixes. Leave with a playbook for choosing useful recurring tasks, improving test feedback, setting token budgets, and deciding when engineers should step in.
Mezzanine Stage
As models become more capable, agent architectures are changing. We’ll trace the evolution from simple loops to collaborating agents, examine the design choices and model advancements behind that shift, and take a deep dive into what comes next.
Outdoor Stage
Stack Overflow’s VP of Product Alex Lato will share insights and practical learnings from bootstrapping and building with Codex. She’ll cover how two members of the Stack technical team created Stack Overflow for Agents, their API-first knowledge exchange, and how her team used the same principles to rapidly reimagine organizational knowledge with their enterprise product Stack Internal. Leave with lessons on using AI to accelerate R&D, modernize existing technology, and bring ambitious new products to market with small teams.
11:30 AM
LunchWalk In The Park
12:00 PM
Festival Pavillion
Join Peter Steinberger for a look at how agents can change the way open-source teams build. Explore his team’s workflow for running agents on a shared server, giving them context across projects, and bringing external contributors into isolated environments. Through live demos, see how cloud sessions and computer use expand what agents can do. Leave with practical ideas for helping a small team build and collaborate at a larger scale.
Cowell Theater
Learn what it takes to run the Codex harness as a service through the Agents API. We’ll explore how agents work in your environment, how applications follow their activity, and how long-running tasks keep going. You’ll leave knowing which infrastructure OpenAI manages and where to focus your own engineering work.
Mezzanine Stage
ChatGPT enables people and teams to take on work that once felt out of reach. Step inside workflows from OpenAI’s power users. Follow a launch as it unfolds, explore competitive research, and see how our Chief People Officer turns expertise into tools. Leave knowing how to move faster, think bigger, and do more of your best work with ChatGPT.
12:45 PM
Cowell Theater
With Astra, ChatGPT Sites, and WebMCP, builders can now build new kinds of interfaces. Learn how to turn an idea into an interactive experience, and how you can collaborate with colleagues and agents to build the next generation of web apps.
Outdoor Stage
Get more successful work from your AI budget. Drawing on customer deployments, this panel explores how to measure cost per completed task, choose the right model and reasoning effort, and reduce spend with prompt caching, Programmatic Tool Calling, and the Batch API. Learn how to spot wasted tokens, account for retries and human rework, and use evals to check that savings preserve quality. Leave with a practical approach to improving the cost of your own application.
1:00 PM
Festival Pavillion
How did games shape AI? What makes a game fun? From Alan Turing to Astra, we’ll explore how games shaped modern AI. Then we’ll learn to build and playtest a retro-style 2D game with Codex, designing its characters, controls, and levels. You’ll see how AI can bring your ideas to life, and how your own taste and creativity make a game worth playing.
1:30 PM
Mezzanine Stage
Learn to use OpenAI’s latest cyber models and open-source CLI and SDK to find and fix vulnerabilities. Drawing on their experience securing OpenAI’s own codebase, Ian Webster and Kyle Brown will show how to scan large codebases, catch vulnerabilities in pull requests, and verify fixes.
Outdoor Stage
Higgsfield CEO and co-founder Alex Mashrabov explains how an agent powered by OpenAI models chooses the right model for each task and provisions compute to run it. Learn when to allocate dedicated GPUs versus call a hosted model, how to deliver predictable throughput for enterprise workloads, and what fails when agents spend real compute on outputs users judge in seconds.
2:15 PM
Festival Pavillion
Make ChatGPT the platform for what you’re building, whether it’s a personal tool, a team workflow, or a product for your users. Learn how plugin extensions bring your tools, knowledge, and interfaces natively into ChatGPT and Codex, and make them easily discoverable. Through live demos and a look at how OpenAI is building, see how to turn an idea into a useful experience, share it with your team, and reach new users.
Cowell Theater
As we move from delegating individual tasks to letting Codex tackle entire problems and goals on its own, it’s possible to harness the power of Codex wherever and however you work. Join us for a session packed with live demos covering what’s new in Codex in the CLI, app, and on the web, and find your new favorite way to build.
Mezzanine Stage
Tejal Patwardhan from the agents post-training team will share how OpenAI researchers are making models more useful and aligned, from computer use to agentic coding to complex knowledge work.
3:00 PM
Cowell Theater
Follow a 2am outage from the first Slack messages to a reviewed fix and an update for the team. Jason shows how voice, Appshots, plugins, and meeting notes help your persistent personal agent work through the investigation. Includes a preview of one of the DevDay launches.
Mezzanine Stage
See GPT-Live in action that can brainstorm, play games, and create images while coordinating work in the background. Then go under the hood to learn how to add voice to your own agents, manage delegation, and design for latency, interruptions, and changing requests.
Outdoor Stage
Qodo CEO Itamar Friedman shares how his team builds with OpenAI to review AI-generated code. Explore the harness behind an adversarial reviewer designed to challenge a coding agent’s assumptions, reasoning, and output—from gathering organizational context to adversarial checks and validation workflows. Learn from the architectures, evals, tradeoffs, and failure modes shaping the system, including how to recognize when model confidence needs human judgment.
4:00 PM
Festival Pavillion
Close out DevDay with Sam Altman, Tejal Patwardhan, and Thibault Sottiaux as they take questions from the community, share their perspectives on the day’s announcements, and discuss what’s next for AI and the developers building with it.
4:45 PM
ReceptionWalk In The Park