您可以根據業務和技術限制,以及對 Vertex AI 和 Google Cloud 生態系統的熟悉程度,為應用程式選取合適的 API 供應商。大多數剛開始整合 Gemini Pro 或 Gemini Flash 的 Android 開發人員,應先從 Gemini Developer API 開始。您可以透過變更模型建構函式中的參數,切換供應商:
Kotlin
// For Vertex AI, use `backend = GenerativeBackend.vertexAI()`valmodel=Firebase.ai(backend=GenerativeBackend.googleAI()).generativeModel("gemini-2.5-flash")valresponse=model.generateContent("Write a story about a magic backpack");valoutput=response.text
Java
// For Vertex AI, use `backend = GenerativeBackend.vertexAI()`GenerativeModelfirebaseAI=FirebaseAI.getInstance(GenerativeBackend.googleAI()).generativeModel("gemini-2.5-flash");// Use the GenerativeModelFutures Java compatibility layer which offers// support for ListenableFuture and Publisher APIsGenerativeModelFuturesmodel=GenerativeModelFutures.from(firebaseAI);Contentprompt=newContent.Builder().addText("Write a story about a magic backpack.").build();ListenableFuture<GenerateContentResponse>response=model.generateContent(prompt);Futures.addCallback(response,newFutureCallback<GenerateContentResponse>(){@OverridepublicvoidonSuccess(GenerateContentResponseresult){StringresultText=result.getText();[...]}@OverridepublicvoidonFailure(Throwablet){t.printStackTrace();}},executor);
[[["容易理解","easyToUnderstand","thumb-up"],["確實解決了我的問題","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["缺少我需要的資訊","missingTheInformationINeed","thumb-down"],["過於複雜/步驟過多","tooComplicatedTooManySteps","thumb-down"],["過時","outOfDate","thumb-down"],["翻譯問題","translationIssue","thumb-down"],["示例/程式碼問題","samplesCodeIssue","thumb-down"],["其他","otherDown","thumb-down"]],["上次更新時間:2025-08-17 (世界標準時間)。"],[],[],null,["# Gemini AI models\n\nThe Gemini Pro and Gemini Flash model families offer Android developers\nmultimodal AI capabilities, running inference in the cloud and processing image,\naudio, video, and text inputs in Android apps.\n\n- **Gemini Pro**: Gemini 2.5 Pro is Google's state-of-the-art thinking model, capable of reasoning over complex problems in code, math, and STEM, as well as analyzing large datasets, codebases, and documents using long context.\n- **Gemini Flash**: The Gemini Flash models deliver next-gen features and improved capabilities, including superior speed, built-in tool use, and a 1M token context window.\n\n| **Note:** This document covers the cloud-based Gemini AI models. For on-device inference, [check out the Gemini Nano documentation](/ai/gemini-nano).\n\nFirebase AI Logic\n-----------------\n\nFirebase AI Logic enables developers to securely and directly add Google's\ngenerative AI into their apps simplifying development, and offers tools and\nproduct integrations for successful production readiness. It provides client\nAndroid SDKs to directly integrate and call Gemini APIs from client code,\nsimplifying development by eliminating the need for a backend.\n\nAPI providers\n-------------\n\nFirebase AI Logic lets you use the following Google Gemini API providers:\nGemini *Developer API* and Vertex *AI Gemini API*.\n**Figure 1.** Firebase AI Logic integration architecture.\n\nHere are the primary differences for each API provider:\n\n[**Gemini Developer API**](/ai/gemini/developer-api):\n\n- Get started at no-cost with a generous free tier without payment information required.\n- Optionally upgrade to the paid tier of the Gemini Developer API to scale as your user base grows.\n- Iterate and experiment with prompts and even get code snippets using [Google AI Studio](https://aistudio.google.com/).\n\n[**Vertex AI Gemini API**](/ai/vertex-ai-firebase):\n\n- Granular control over [where you access the model](https://cloud.google.com/compute/docs/regions-zones).\n- Ideal for developers already embedded in the Vertex AI/Google Cloud ecosystem.\n- Iterate and experiment with prompts and even get code snippets using [Vertex AI Studio](https://cloud.google.com/vertex-ai/generative-ai/docs/start/quickstarts/quickstart).\n\nSelecting the appropriate API provider for your application is based on your\nbusiness and technical constraints, and familiarity with the Vertex AI and\nGoogle Cloud ecosystem. Most Android developers just getting started with Gemini\nPro or Gemini Flash integrations should begin with the Gemini Developer API.\nSwitching between providers is done by changing the parameter in the model\nconstructor: \n\n### Kotlin\n\n // For Vertex AI, use `backend = GenerativeBackend.vertexAI()`\n val model = Firebase.ai(backend = GenerativeBackend.googleAI())\n .generativeModel(\"gemini-2.5-flash\")\n\n val response = model.generateContent(\"Write a story about a magic backpack\");\n val output = response.text\n\n### Java\n\n // For Vertex AI, use `backend = GenerativeBackend.vertexAI()`\n GenerativeModel firebaseAI = FirebaseAI.getInstance(GenerativeBackend.googleAI())\n .generativeModel(\"gemini-2.5-flash\");\n\n // Use the GenerativeModelFutures Java compatibility layer which offers\n // support for ListenableFuture and Publisher APIs\n GenerativeModelFutures model = GenerativeModelFutures.from(firebaseAI);\n\n Content prompt = new Content.Builder()\n .addText(\"Write a story about a magic backpack.\")\n .build();\n\n ListenableFuture\u003cGenerateContentResponse\u003e response = model.generateContent(prompt);\n Futures.addCallback(response, new FutureCallback\u003cGenerateContentResponse\u003e() {\n @Override\n public void onSuccess(GenerateContentResponse result) {\n String resultText = result.getText();\n [...]\n }\n\n @Override\n public void onFailure(Throwable t) {\n t.printStackTrace();\n }\n }, executor);\n\nSee the full [list of available generative AI models](https://firebase.google.com/docs/vertex-ai/models) supported\nby Firebase AI Logic client SDKs.\n\nFirebase services\n-----------------\n\nIn addition to access to the Gemini API, Firebase AI Logic offers a set of\nservices to simplify the deployment of AI-enabled features to your app and get\nready for production:\n\n### App Check\n\n[Firebase App Check](https://firebase.google.com/docs/app-check) safeguards app backends from abuse by\nensuring only authorized clients access resources. It integrates with Google\nservices (including Firebase and Google Cloud) and custom backends. App Check\nuses [Play Integrity](/google/play/integrity) to verify that requests originate from the authentic\napp and an untampered device.\n\n### Remote Config\n\nInstead of hardcoding the model name in your app, we recommend using a\nserver-controlled variable using [Firebase Remote Config](https://firebase.google.com/docs/remote-config). This\nlets you dynamically update the model your app uses without having to deploy a\nnew version of your app or require your users to pick up a new version. You can\nalso use Remote Config to [A/B test](https://firebase.google.com/docs/ab-testing/abtest-config) models and prompts.\n\n### AI monitoring\n\nTo understand how your AI-enabled features are performing you can use the [AI\nmonitoring dashboard](https://firebase.google.com/docs/vertex-ai/monitoring) within the Firebase console. You'll get\nvaluable insights into usage patterns, performance metrics, and debugging\ninformation for your Gemini API calls.\n\nMigrate to Firebase AI Logic\n----------------------------\n\nIf you're already using the Vertex AI in Firebase SDK in your app, read the\n[migration guide](https://firebase.google.com/docs/vertex-ai/migrate-to-latest-sdk)."]]