We're excited to announce our first challenge with Google AI!
Running through September 14, the Google AI Studio Multimodal Challenge invites you to build and deploy innovative applets that showcase the incredible power of Gemini's multimodal capabilities.
There's one prompt for this challenge but three chances to win! We hope you give it a try.
Our Prompt
Your mission is to build and deploy an applet on Google AI Studio that shows off the power of Gemini in multimodal content understanding and/or generation.
Your applet must meet these requirements:
- Built on Google AI Studio
- Deployed using Cloud Run
- Utilize at least one of the below multimodal functionality:
- Gemini 2.5 Pro / Flash-supported image, video, or audio understanding
- Gemini 2.5 Flash Image (free tier available for Sept 6-7)
- Live API (free tier supports 3 concurrent sessions)
- Imagen (requires paid tier)
- Veo (requires paid tier)
Think interactive storytelling experiences, tools that process audio and visual content, platforms that understand and generate content across multiple media types. If you can think it, you can build it!
Important Notes:
- Use of paid APIs are purely optional and won't enhance your chance of winning.
- If you used Gemini 2.5 Flash Image during the free trial period and it is no longer available for your app during the judging period, be sure to include a video so we can see your project in action.
- "Build Apps with Google AI Studio" DEV Education Track submissions do not count towards this challenge, you must build and deploy a new, original applet.
Prizes
We'll select three winners for this challenge. Each winner will receive:
- $1,000 USD
- DEV++ Membership
- Exclusive DEV Badge
All Participants with a valid submission will receive a completion badge on their DEV profile.
Getting Started with Google AI Studio π οΈ
New to Google AI Studio? You can get to know Google AI Studio by tackling the "Build Apps with Google AI Studio" DEV Education Track:
DEV Tutorial
Eligibility Requirements
This specific challenge includes the following restrictions:
Territories and regions excluded:
Afghanistan, Belarus, Central African Republic, Cuba, Equatorial Guinea, Iran, Iraq, Kosovo, Libya, Myanmar (Burma), North Korea, Russia, South Sudan, Sudan, Syria, Tanzania, Venezuela, Yemen,
Crimea, the so-called Donetsk People's Republic (DNR), the so-called Luhansk People's Republic (LNR).Additional participant exclusions for this challenge: Government institution employees, Google employees
Open only to 18+
These challenge-specific restrictions are in addition to all eligibility requirements and exclusions detailed in our General Contest Official Rules and Google AI Studio Multimodal v25.09.03 Contest Rules.
Important Dates
- September 3: Google AI Studio Multimodal Challenge begins!
- September 14: Submissions due at 11:59 PM PDT
- October 2: Winners Announced
We can't wait to see the innovative applets you build with Google AI Studio! Questions about the challenge? Ask them below.
Good luck and happy coding!
Top comments (25)
This is going to be epic
Can't wait to see these
Thats cool
This looks awesome
Interesting challenge, but I have a question: If I recall correctly, doesn't deploying an applet to Cloud Run require a billing account, which in turn incurs billing for the participant per run? Is the deployment part mandatory?
Hey @delaaja, great call out. Deployment is required, and Cloud Run actually has a substantial free tier, which should cover anything incurred from our judging usage. Please see this page here to see what's covered, so you can be aware while testing (pre-submission): https://cloud.google.com/run?e=48754805&hl=en#pricing
Exactly!
To be honest, I don't really understand the point of the challenge. The models that can actually generate something useful are all paid. Fine, I added billing. Now Iβve built a game, and it generates images during gameplay. The limit for the first tier is 70 requests per day. I have 4 images per run, so basically I used up all my requests just this morning and now I canβt even record a video of the game. Iβll have to wait until tomorrow.
The third model, imagen-3.0-generate-001, isnβt even available via API, it says itβs not supported. And to move to Tier 2, I apparently need to spend $300 somehow.
So overall, itβs not really clear what the organizers were expecting. I'll wait for tomorrow to record my demo videos but overall so far experience is not what I expected for the goal "Think interactive storytelling experiences, tools that process audio and visual content, platforms that understand and generate content across multiple media types."
ok, I found that gemini-2.5-flash-image-preview also can generate images - but Google AI Studio doesn't know how to use this model so in the end I fixed the code by myself.
Gonna be felicitating!
This looks fun!!!
Exited!
Hey, this looks interesting. I'm experimenting a bit with AI studio, and initial results look promising. But as I evolve the app, I'm noticing that it's breaking things that were previously working. Should I ask it to write tests, or can I in some other way entice it to keep previous instructions into account, or ask me if anything seems contradictory?
Thinking about this a bit further, I suspect that this is a cxontext issue, where it just can't consider the whole context as I ask more and more iterations. Maybe a tool that compacts the conversation in AI studio would be good, to summarize to the agent what we already have.
Some comments may only be visible to logged-in visitors. Sign in to view all comments.