OpenAI Codex Hackathon — Berkeley — Kickoff
Hosted on LumaFetched about 4 hours ago
Tuesday, March 17, 2026
to Wednesday, March 18, 2026
•
2 hours long
Artificial IntelligenceMachine LearningOpen SourceInnovation
Student only
This hackathons is only open to students. Double check the event page for more information as this may mean only those from a particular university/country are eligible.
Event Type
in person
32
Participants
$15,000
Prize Pool
2
Est. Projects
Organizers
OpenAI Codex Hackathon — Berkeley
Haas AI Club presents Berkeley’s first Codex community hackathon
Join us for a day of building with Codex and UC Berkeley Haas AI Club.
Build something new from scratch using Codex. Run parallel tasks, experiment with Skills, and use worktrees to operate at a higher level of speed and ambition.
This is a hands-on builder event, the short time frame is designed to keep the room focused on shipping at current speed of software.
Who This Event Is For
This event is designed for builders who are curious about Codex and agentic software engineering.
This is the event kickoff, post which individuals can keep hacking till midnight.
Submission
3-min demo video submissions are due by midnight on a link that'll be provided. Judges will grade the video & we'll share it with Haas AI community and OpenAI.
What do I build:
Build something meaningful with Codex — either:
Build a new project from scratch that can turn into a venture fundable startup. Prototype a new breakthrough developer tool or agentic workflow.
This is an individual event. Each builder will need to share a YC-style demo video showing what they've built, how they did it and the problem it addresses.
To keep the event focused, attendance is capped.
If any of these ideas excite you — apply.
🔹 Agentic Workflows
Build tools that maximise leverage from Codex as an AI coding agent in whatever area of work drains you or customers down.
🔹 UX for Agentic Applications
Design innovative interfaces or workflows for working with agentic software.
🔹 Multimodal Intelligence
Create applications that reason across multiple modalities.
🔹 Domain Agents
Build agents that operate within real-world constraints as domain experts.
🔹 Building Evals
Develop tools to measure, debug, and systematically improve agent performance.
Prizes
🥇 1st Place — $15,000 USD in OpenAI API credits
🥈 2nd Place — $10,000 USD in OpenAI API credits
🥉 3rd Place — $5,000 USD in OpenAI API credits
Others - Pizza 🍕, Drinks 🥤 and learning to build using the best agentic coding models.
Kickoff Details
📅 Tuesday, March 17th
🕘 2pm ~ 5pm (Submission until Midnight)
📍 N470 - Chou Hall
Judging criteria
Projects will be scored across four categories:
Innovation & Frontier AI – 30 points.
Market Potential – 30 points.
Technical Execution & Reliability – 20 points.
Presentation & Communication – 20 points.
Registration
Due to venue capacity, registration is subject to approval.
Please complete the registration form carefully. Incomplete or inaccurate submissions may not be admitted.