gpt 5 o gpt 5 mini

infra/main.bicep

@@ -91,6 +91,30 @@ param privateEndpointVNetPrefix string = '192.168.0.0/16'
 param privateEndpointSubnetAddressPrefix string = '192.168.0.0/24'
 param appServiceBackendSubnetAddressPrefix string = '192.168.1.0/24'
 
+@description('Additional OpenAI deployments to create if you want more to choose from(e.g., gpt-5, gpt-5-mini)')
+param additionalLlmDeployments array = [
+  {
+    name: 'gpt-5'
+    model: {
+      name: 'gpt-5'
+      version: '2025-08-07'
+    }
+    sku: {
+      capacity: 120
+    }
+  }
+  {
+    name: 'gpt-5-mini'
+    model: {
+      name: 'gpt-5-mini'
+      version: '2025-08-07'
+    }
+    sku: {
+      capacity: 120
+    }
+  }
+]
+
 var resourceToken = toLower(uniqueString(subscription().id, name, location))
 var tags = { 'azd-env-name': name }
 
@@ -134,6 +158,7 @@ module resources 'resources.bicep' = {
     privateEndpointVNetPrefix: privateEndpointVNetPrefix
     privateEndpointSubnetAddressPrefix: privateEndpointSubnetAddressPrefix
     appServiceBackendSubnetAddressPrefix: appServiceBackendSubnetAddressPrefix
+    additionalLlmDeployments: additionalLlmDeployments
   }
 }
 

infra/resources.bicep

@@ -43,6 +43,9 @@ param privateEndpointVNetPrefix string = '192.168.0.0/16'
 param privateEndpointSubnetAddressPrefix string = '192.168.0.0/24'
 param appServiceBackendSubnetAddressPrefix string = '192.168.1.0/24'
 
+@description('Optional additional LLM deployments to create on the same Azure OpenAI resource')
+param additionalLlmDeployments array = []
+
 var openai_name = toLower('${name}-aillm-${resourceToken}')
 var openai_dalle_name = toLower('${name}-aidalle-${resourceToken}')
 
@@ -73,7 +76,7 @@ var databaseName = 'chat'
 var historyContainerName = 'history'
 var configContainerName = 'config'
 
-var llmDeployments = [
+var baseLlmDeployments = [
   {
     name: chatGptDeploymentName
     model: {
@@ -97,6 +100,23 @@ var llmDeployments = [
   }
 ]
 
+var mappedAdditionalDeployments = [
+  for addition in additionalLlmDeployments: {
+    name: addition.name
+    model: {
+      format: 'OpenAI'
+      name: addition.model.name
+      version: addition.model.version
+    }
+    sku: {
+      name: 'GlobalStandard'
+      capacity: addition.capacity
+    }
+  }
+]
+
+var llmDeployments = concat(baseLlmDeployments, mappedAdditionalDeployments)
+
 module privateEndpoints 'private_endpoints_core.bicep' = if (usePrivateEndpoints) {
   name: 'private-endpoints'
   params: {
@@ -148,6 +168,10 @@ var appSettingsCommon = [
     name: 'SCM_DO_BUILD_DURING_DEPLOYMENT'
     value: 'true'
   }
+  {
+    name: 'AZURE_OPENAI_API_EMBEDDINGS_DEPLOYMENT_NAME'
+    value: embeddingDeploymentName
+  }
   {
     name: 'AZURE_OPENAI_API_INSTANCE_NAME'
     value: openai_name
@@ -156,10 +180,6 @@ var appSettingsCommon = [
     name: 'AZURE_OPENAI_API_DEPLOYMENT_NAME'
     value: chatGptDeploymentName
   }
-  {
-    name: 'AZURE_OPENAI_API_EMBEDDINGS_DEPLOYMENT_NAME'
-    value: embeddingDeploymentName
-  }
   {
     name: 'AZURE_OPENAI_API_VERSION'
     value: openai_api_version
@@ -214,6 +234,13 @@ var appSettingsCommon = [
   }
 ]
 
+var additionalModelSettingsDeployment = [
+  for (deployment, i) in mappedAdditionalDeployments: {
+    name: 'AZURE_OPENAI_API_DEPLOYMENT_NAME_MODEL_${i + 1}'
+    value: deployment.model.name
+  }
+]
+
 var appSettingsWithLocalAuth = disableLocalAuth
   ? []
   : [
@@ -258,7 +285,7 @@ resource webApp 'Microsoft.Web/sites@2024-04-01' = {
       appCommandLine: 'next start'
       ftpsState: 'Disabled'
       minTlsVersion: '1.2'
-      appSettings: concat(appSettingsCommon, appSettingsWithLocalAuth)
+      appSettings: concat(appSettingsCommon, additionalModelSettingsDeployment, appSettingsWithLocalAuth)
     }
   }
   identity: { type: 'SystemAssigned' }
@@ -503,12 +530,10 @@ resource llmdeployment 'Microsoft.CognitiveServices/accounts/deployments@2023-05
       model: deployment.model
       /*raiPolicyName: contains(deployment, 'raiPolicyName') ? deployment.raiPolicyName : null*/
     }
-    sku: contains(deployment, 'sku')
-      ? deployment.sku
-      : {
-          name: 'Standard'
-          capacity: deployment.capacity
-        }
+    sku: deployment.?sku ?? {
+      name: 'Standard'
+      capacity: deployment.capacity
+    }
   }
 ]
 

src/features/chat-page/chat-header/chat-header.tsx

@@ -6,6 +6,14 @@ import { ChatDocumentModel, ChatThreadModel } from "../chat-services/models";
 import { DocumentDetail } from "./document-detail";
 import { ExtensionDetail } from "./extension-detail";
 import { PersonaDetail } from "./persona-detail";
+import { chatStore, useChat } from "@/features/chat-page/chat-store";
+import {
+  Select,
+  SelectContent,
+  SelectItem,
+  SelectTrigger,
+  SelectValue,
+} from "@/features/ui/select";
 
 interface Props {
   chatThread: ChatThreadModel;
@@ -29,7 +37,9 @@ export const ChatHeader: FC<Props> = (props) => {
             {persona}
           </span>
         </div>
-        <div className="flex gap-2">
+        <div className="flex gap-2 items-center">
+          {/* Model selector */}
+          <ModelSelect />
           <PersonaDetail chatThread={props.chatThread} />
           <DocumentDetail chatDocuments={props.chatDocuments} />
           <ExtensionDetail
@@ -43,3 +53,22 @@ export const ChatHeader: FC<Props> = (props) => {
     </div>
   );
 };
+
+const ModelSelect: FC = () => {
+  const { selectedModel } = useChat();
+  const onChange = (v: string) => chatStore.updateSelectedModel(v);
+
+  return (
+    <Select value={selectedModel} onValueChange={onChange}>
+      <SelectTrigger className="w-44">
+        <SelectValue placeholder="Choose model" />
+      </SelectTrigger>
+      <SelectContent>
+        {/* Use your Azure OpenAI deployment names here */}
+        <SelectItem value="gpt-4o">gpt-4o</SelectItem>
+        <SelectItem value="gpt-5">gpt-5</SelectItem>
+        <SelectItem value="gpt-5-mini">gpt-5-mini</SelectItem>
+      </SelectContent>
+    </Select>
+  );
+};

src/features/chat-page/chat-services/chat-api/chat-api-extension.ts

@@ -13,10 +13,11 @@ export const ChatApiExtensions = async (props: {
   history: ChatCompletionMessageParam[];
   extensions: RunnableToolFunction<any>[];
   signal: AbortSignal;
+  model?: string;
 }): Promise<ChatCompletionStreamingRunner> => {
-  const { userMessage, history, signal, chatThread, extensions } = props;
+  const { userMessage, history, signal, chatThread, extensions, model } = props;
 
-  const openAI = OpenAIInstance();
+  const openAI = OpenAIInstance(model);
   const systemMessage = await extensionsSystemMessage(chatThread);
   return openAI.beta.chat.completions.runTools(
     {

src/features/chat-page/chat-services/chat-api/chat-api-multimodal.tsx

@@ -9,10 +9,11 @@ export const ChatApiMultimodal = (props: {
   userMessage: string;
   file: string;
   signal: AbortSignal;
+  model?: string;
 }): ChatCompletionStreamingRunner => {
-  const { chatThread, userMessage, signal, file } = props;
+  const { chatThread, userMessage, signal, file, model } = props;
 
-  const openAI = OpenAIInstance();
+  const openAI = OpenAIInstance(model);
 
   return openAI.beta.chat.completions.stream(
     {

src/features/chat-page/chat-services/chat-api/chat-api-rag.ts

@@ -17,10 +17,11 @@ export const ChatApiRAG = async (props: {
   userMessage: string;
   history: ChatCompletionMessageParam[];
   signal: AbortSignal;
+  model?: string;
 }): Promise<ChatCompletionStreamingRunner> => {
-  const { chatThread, userMessage, history, signal } = props;
+  const { chatThread, userMessage, history, signal, model } = props;
 
-  const openAI = OpenAIInstance();
+  const openAI = OpenAIInstance(model);
 
   const documentResponse = await SimilaritySearch(
     userMessage,

src/features/chat-page/chat-services/chat-api/chat-api.ts

@@ -72,6 +72,7 @@ export const ChatAPIEntry = async (props: UserPrompt, signal: AbortSignal) => {
         userMessage: props.message,
         history: history,
         signal: signal,
+        model: props.model,
       });
       break;
     case "multimodal":
@@ -80,6 +81,7 @@ export const ChatAPIEntry = async (props: UserPrompt, signal: AbortSignal) => {
         userMessage: props.message,
         file: props.multimodalImage,
         signal: signal,
+        model: props.model,
       });
       break;
     case "extensions":
@@ -89,6 +91,7 @@ export const ChatAPIEntry = async (props: UserPrompt, signal: AbortSignal) => {
         history: history,
         extensions: extension,
         signal: signal,
+        model: props.model,
       });
       break;
   }

src/features/chat-page/chat-services/models.ts

@@ -40,6 +40,7 @@ export interface UserPrompt {
   id: string; // thread id
   message: string;
   multimodalImage: string;
+  model?: string;
 }
 
 export interface ChatDocumentModel {

src/features/chat-page/chat-store.tsx

@@ -35,6 +35,7 @@ class ChatState {
   public autoScroll: boolean = false;
   public userName: string = "";
   public chatThreadId: string = "";
+  public selectedModel: string = "";
 
   private chatThread: ChatThreadModel | undefined;
 
@@ -58,6 +59,10 @@ class ChatState {
     this.loading = value;
   }
 
+  public updateSelectedModel(model: string) {
+    this.selectedModel = model;
+  }
+
   public initChatSession({
     userName,
     messages,
@@ -284,6 +289,7 @@ class ChatState {
     const body = JSON.stringify({
       id: this.chatThreadId,
       message: this.input,
+      model: this.selectedModel,
     });
     formData.append("content", body);
 

src/features/common/services/openai.ts

@@ -4,28 +4,46 @@ import { AzureOpenAI } from "openai";
 
 const USE_MANAGED_IDENTITIES = process.env.USE_MANAGED_IDENTITIES === "true";
 
-export const OpenAIInstance =  () => {
+export const OpenAIInstance =  (deploymentOverride?: string) => {
   const endpointSuffix = process.env.AZURE_OPENAI_API_ENDPOINT_SUFFIX || "openai.azure.com";
-  let token = process.env.AZURE_OPENAI_API_KEY;
+  const token = process.env.AZURE_OPENAI_API_KEY;
+
+  // Resolve model/instance/deployment based on optional override
+  const instanceName = process.env.AZURE_OPENAI_API_INSTANCE_NAME;
+  const defaultDeployment = process.env.AZURE_OPENAI_API_DEPLOYMENT_NAME;
+  const apiVersion = process.env.AZURE_OPENAI_API_VERSION;
+
+  let deploymentName = defaultDeployment;
+
+  if (deploymentOverride) {
+    for (let i = 1; i <= 100; i++) {
+      const key = `AZURE_OPENAI_API_DEPLOYMENT_NAME_MODEL_${i}`;
+      const instanceCandidate = process.env[key];
+      if (!instanceCandidate) continue;
+      if (instanceCandidate === deploymentOverride) {
+        deploymentName = instanceCandidate;
+        break;
+      }
+    }
+  }
+
   if (USE_MANAGED_IDENTITIES) {
     const credential = new DefaultAzureCredential();
     const scope = "https://cognitiveservices.azure.com/.default";
     const azureADTokenProvider = getBearerTokenProvider(credential, scope);
-    const deployment = process.env.AZURE_OPENAI_API_DEPLOYMENT_NAME;
-    const apiVersion = process.env.AZURE_OPENAI_API_VERSION;
     const client = new AzureOpenAI({
       azureADTokenProvider,
-      deployment,
+      deployment: deploymentName,
       apiVersion,
-      baseURL: `https://${process.env.AZURE_OPENAI_API_INSTANCE_NAME}.${endpointSuffix}/openai/deployments/${process.env.AZURE_OPENAI_API_DEPLOYMENT_NAME}`
+      baseURL: `https://${instanceName}.${endpointSuffix}/openai/deployments/${deploymentName}`
     });
     return client;
   } else {
     const openai = new OpenAI({
       apiKey: token,
-      baseURL: `https://${process.env.AZURE_OPENAI_API_INSTANCE_NAME}.${endpointSuffix}/openai/deployments/${process.env.AZURE_OPENAI_API_DEPLOYMENT_NAME}`,
-      defaultQuery: { "api-version": process.env.AZURE_OPENAI_API_VERSION },
-      defaultHeaders: { "api-key": process.env.AZURE_OPENAI_API_KEY },
+      baseURL: `https://${instanceName}.${endpointSuffix}/openai/deployments/${deploymentName}`,
+      defaultQuery: { "api-version": apiVersion },
+      defaultHeaders: { "api-key": token },
     });
     return openai;
   }