Posted by Thomas Ezan, Senior Developer Relations Engineer
.png)
Gemini 3 optimized for low-latency
Gemini 3 is our most clever mannequin household up to now. With the launch of Gemini 3 Flash, we’re making that intelligence extra accessible for low-latency and cost-effective use circumstances. Whereas Gemini 3 Professional is designed for advanced reasoning, Gemini 3 Flash is engineered to be considerably sooner and less expensive on your manufacturing apps.
Seamless integration with Firebase AI Logic
Similar to the Professional mannequin, Gemini 3 Flash is out there in preview instantly via the Firebase AI Logic SDK. This implies you’ll be able to combine it into your Android app without having to do any advanced server facet setup.
Right here is how you can add it to your Kotlin code:
val mannequin = Firebase.ai(backend = GenerativeBackend.googleAI())
.generativeModel(
modelName = "gemini-3-flash-preview")Scale with Confidence
As well as, Firebase lets you maintain your progress safe and manageable with:
AI Monitoring
The Firebase AI monitoring dashboard provides you visibility into latency, success charges, and prices, permitting you to slice knowledge by mannequin identify to see precisely how the mannequin performs.
Server Immediate Templates
You should utilize server immediate templates to retailer your immediate and schema securely on Firebase servers as a substitute of hardcoding them in your app binary. This functionality ensures your delicate prompts stay safe, prevents unauthorized immediate extraction, and permits for sooner iteration with out requiring app updates.
--- mannequin: 'gemini-3-flash-preview' enter: schema: subject: kind: 'string' minLength: 2 maxLength: 40 size: kind: 'quantity' minimal: 1 most: 200 language: kind: 'string' --- {{position "system"}} You're a storyteller that tells good and joyful tales with joyful endings. {{position "person"}} Create a narrative about {{subject}} with the size of {{size}} phrases within the {{language}} language.
Immediate template outlined on the Firebase Console
val generativeModel = Firebase.ai.templateGenerativeModel() val response = generativeModel.generateContent("storyteller-v10", mapOf( "subject" to subject, "size" to size, "language" to language ) ) _output.worth = response.textual content
Code snippet to entry to the immediate template
Gemini 3 Flash for AI improvement help in Android Studio
Gemini 3 Flash can also be obtainable for AI help in Android Studio. Whereas Gemini 3 Professional Preview is our greatest mannequin for coding and agentic experiences, Gemini 3 Flash is engineered for velocity, and nice for widespread improvement duties and questions.
The brand new mannequin is rolling out to builders utilizing Gemini in Android Studio at no-cost (default mannequin) beginning at the moment. For greater utilization price limits and longer periods with Agent Mode, you’ll be able to use an AI Studio API key to leverage the total capabilities of both Gemini 3 Flash or Gemini 3 Professional. We’re additionally rolling out Gemini 3 mannequin household entry with greater utilization price limits to builders who’ve Gemini Code Help Normal or Enterprise licenses. Your IT administrator might want to allow entry to preview fashions via the Google Cloud console.
Get Began Immediately
You can begin experimenting with Gemini 3 Flash through Firebase AI Logic at the moment. Study extra about it within the Android and Firebase documentation. Check out any of the brand new Gemini 3 fashions in Android Studio for improvement help, and tell us what you suppose! As all the time you’ll be able to comply with us throughout LinkedIn, Weblog, YouTube, and X.