diff --git a/docs/samples/Microsoft.ML.GenAI.Samples/MEAI/Llama3_1.cs b/docs/samples/Microsoft.ML.GenAI.Samples/MEAI/Llama3_1.cs
index 4416fe757c..7fe73e72f4 100644
--- a/docs/samples/Microsoft.ML.GenAI.Samples/MEAI/Llama3_1.cs
+++ b/docs/samples/Microsoft.ML.GenAI.Samples/MEAI/Llama3_1.cs
@@ -46,7 +46,7 @@ public static async Task RunAsync(string weightFolder, string checkPointName = "
             """;
         var chatMessage = new ChatMessage(ChatRole.User, task);
 
-        await foreach (var response in client.CompleteStreamingAsync([chatMessage]))
+        await foreach (var response in client.GetStreamingResponseAsync([chatMessage]))
         {
             Console.Write(response.Text);
         }
diff --git a/docs/samples/Microsoft.ML.GenAI.Samples/MEAI/Phi3.cs b/docs/samples/Microsoft.ML.GenAI.Samples/MEAI/Phi3.cs
index 7bba28bc14..a10c01dfb3 100644
--- a/docs/samples/Microsoft.ML.GenAI.Samples/MEAI/Phi3.cs
+++ b/docs/samples/Microsoft.ML.GenAI.Samples/MEAI/Phi3.cs
@@ -36,7 +36,7 @@ public static async Task RunAsync(string weightFolder)
             """;
         var chatMessage = new ChatMessage(ChatRole.User, task);
 
-        await foreach (var response in client.CompleteStreamingAsync([chatMessage]))
+        await foreach (var response in client.GetStreamingResponseAsync([chatMessage]))
         {
             Console.Write(response.Text);
         }
diff --git a/eng/Versions.props b/eng/Versions.props
index 0cc944a243..d2286a30d6 100644
--- a/eng/Versions.props
+++ b/eng/Versions.props
@@ -46,7 +46,7 @@
     <MicrosoftDotNetInteractiveVersion>1.0.0-beta.24375.2</MicrosoftDotNetInteractiveVersion>
     <MicrosoftMLOnnxRuntimeVersion>1.18.1</MicrosoftMLOnnxRuntimeVersion>
     <MlNetMklDepsVersion>0.0.0.12</MlNetMklDepsVersion>
-    <MicrosoftExtensionsAIVersion>9.0.1-preview.1.24570.5</MicrosoftExtensionsAIVersion>
+    <MicrosoftExtensionsAIVersion>9.3.0-preview.1.25114.11</MicrosoftExtensionsAIVersion>
     <!-- runtime.native.System.Data.SqlClient.sni is not updated by dependency flow as it is not produced live anymore. -->
     <RuntimeNativeSystemDataSqlClientSniVersion>4.4.0</RuntimeNativeSystemDataSqlClientSniVersion>
     <!--
diff --git a/src/Microsoft.ML.GenAI.Core/CausalLMPipelineChatClient.cs b/src/Microsoft.ML.GenAI.Core/CausalLMPipelineChatClient.cs
index 7d1ffee34d..9a6c7c222b 100644
--- a/src/Microsoft.ML.GenAI.Core/CausalLMPipelineChatClient.cs
+++ b/src/Microsoft.ML.GenAI.Core/CausalLMPipelineChatClient.cs
@@ -20,6 +20,7 @@ public abstract class CausalLMPipelineChatClient<TTokenizer, TCausalLMModel> : I
 {
     private readonly ICausalLMPipeline<TTokenizer, TCausalLMModel> _pipeline;
     private readonly IMEAIChatTemplateBuilder _chatTemplateBuilder;
+    private readonly ChatClientMetadata _metadata;
 
     public CausalLMPipelineChatClient(
         ICausalLMPipeline<TTokenizer, TCausalLMModel> pipeline,
@@ -27,14 +28,12 @@ public CausalLMPipelineChatClient(
         ChatClientMetadata? metadata = null)
     {
         var classNameWithType = $"{nameof(CausalLMPipelineChatClient<TTokenizer, TCausalLMModel>)}<{typeof(TTokenizer).Name}, {typeof(TCausalLMModel).Name}>";
-        Metadata ??= new ChatClientMetadata(providerName: classNameWithType, modelId: typeof(TCausalLMModel).Name);
+        _metadata = new ChatClientMetadata(providerName: classNameWithType, modelId: typeof(TCausalLMModel).Name);
         _chatTemplateBuilder = chatTemplateBuilder;
         _pipeline = pipeline;
     }
 
-    public ChatClientMetadata Metadata { get; }
-
-    public virtual Task<ChatCompletion> CompleteAsync(IList<ChatMessage> chatMessages, ChatOptions? options = null, CancellationToken cancellationToken = default)
+    public virtual Task<ChatResponse> GetResponseAsync(IList<ChatMessage> chatMessages, ChatOptions? options = null, CancellationToken cancellationToken = default)
     {
         var prompt = _chatTemplateBuilder.BuildPrompt(chatMessages, options);
         var stopSequences = options?.StopSequences ?? Array.Empty<string>();
@@ -46,7 +45,7 @@ public virtual Task<ChatCompletion> CompleteAsync(IList<ChatMessage> chatMessage
             stopSequences: stopSequences.ToArray()) ?? throw new InvalidOperationException("Failed to generate a reply.");
 
         var chatMessage = new ChatMessage(ChatRole.Assistant, output);
-        return Task.FromResult(new ChatCompletion([chatMessage])
+        return Task.FromResult(new ChatResponse([chatMessage])
         {
             CreatedAt = DateTime.UtcNow,
             FinishReason = ChatFinishReason.Stop,
@@ -54,7 +53,7 @@ public virtual Task<ChatCompletion> CompleteAsync(IList<ChatMessage> chatMessage
     }
 
 #pragma warning disable CS1998 // Async method lacks 'await' operators and will run synchronously
-    public virtual async IAsyncEnumerable<StreamingChatCompletionUpdate> CompleteStreamingAsync(
+    public virtual async IAsyncEnumerable<ChatResponseUpdate> GetStreamingResponseAsync(
 #pragma warning restore CS1998 // Async method lacks 'await' operators and will run synchronously
         IList<ChatMessage> chatMessages,
         ChatOptions? options = null,
@@ -69,7 +68,7 @@ public virtual async IAsyncEnumerable<StreamingChatCompletionUpdate> CompleteStr
             temperature: options?.Temperature ?? 0.7f,
             stopSequences: stopSequences.ToArray()))
         {
-            yield return new StreamingChatCompletionUpdate
+            yield return new ChatResponseUpdate
             {
                 Role = ChatRole.Assistant,
                 Text = output,
@@ -83,6 +82,8 @@ public virtual void Dispose()
     }
 
     public virtual object? GetService(Type serviceType, object? serviceKey = null) =>
-        serviceKey is null && serviceType is not null && serviceType.IsAssignableFrom(GetType()) ? this :
+        serviceKey is not null ? null :
+        serviceType == typeof(ChatClientMetadata) ? _metadata :
+        serviceType.IsAssignableFrom(GetType()) ? this :
         null;
 }
diff --git a/src/Microsoft.ML.GenAI.LLaMA/Llama3CausalLMChatClient.cs b/src/Microsoft.ML.GenAI.LLaMA/Llama3CausalLMChatClient.cs
index ad0b58c3be..cc275e7686 100644
--- a/src/Microsoft.ML.GenAI.LLaMA/Llama3CausalLMChatClient.cs
+++ b/src/Microsoft.ML.GenAI.LLaMA/Llama3CausalLMChatClient.cs
@@ -24,7 +24,7 @@ public Llama3CausalLMChatClient(
     {
     }
 
-    public override Task<ChatCompletion> CompleteAsync(
+    public override Task<ChatResponse> GetResponseAsync(
         IList<ChatMessage> chatMessages,
         ChatOptions? options = null,
         CancellationToken cancellationToken = default)
@@ -40,10 +40,10 @@ public override Task<ChatCompletion> CompleteAsync(
             options.StopSequences = new List<string> { _eotToken };
         }
 
-        return base.CompleteAsync(chatMessages, options, cancellationToken);
+        return base.GetResponseAsync(chatMessages, options, cancellationToken);
     }
 
-    public override IAsyncEnumerable<StreamingChatCompletionUpdate> CompleteStreamingAsync(
+    public override IAsyncEnumerable<ChatResponseUpdate> GetStreamingResponseAsync(
         IList<ChatMessage> chatMessages,
         ChatOptions? options = null,
         CancellationToken cancellationToken = default)
@@ -52,6 +52,6 @@ public override IAsyncEnumerable<StreamingChatCompletionUpdate> CompleteStreamin
         options.StopSequences ??= [];
         options.StopSequences.Add(_eotToken);
 
-        return base.CompleteStreamingAsync(chatMessages, options, cancellationToken);
+        return base.GetStreamingResponseAsync(chatMessages, options, cancellationToken);
     }
 }
diff --git a/src/Microsoft.ML.GenAI.Phi/Phi3/Phi3CausalLMChatClient.cs b/src/Microsoft.ML.GenAI.Phi/Phi3/Phi3CausalLMChatClient.cs
index 9477f423b7..d63823ea51 100644
--- a/src/Microsoft.ML.GenAI.Phi/Phi3/Phi3CausalLMChatClient.cs
+++ b/src/Microsoft.ML.GenAI.Phi/Phi3/Phi3CausalLMChatClient.cs
@@ -29,7 +29,7 @@ public Phi3CausalLMChatClient(
     {
     }
 
-    public override Task<ChatCompletion> CompleteAsync(
+    public override Task<ChatResponse> GetResponseAsync(
         IList<ChatMessage> chatMessages,
         ChatOptions? options = null,
         CancellationToken cancellationToken = default)
@@ -45,10 +45,10 @@ public override Task<ChatCompletion> CompleteAsync(
             options.StopSequences = [_eotToken];
         }
 
-        return base.CompleteAsync(chatMessages, options, cancellationToken);
+        return base.GetResponseAsync(chatMessages, options, cancellationToken);
     }
 
-    public override IAsyncEnumerable<StreamingChatCompletionUpdate> CompleteStreamingAsync(
+    public override IAsyncEnumerable<ChatResponseUpdate> GetStreamingResponseAsync(
         IList<ChatMessage> chatMessages,
         ChatOptions? options = null,
         CancellationToken cancellationToken = default)
@@ -57,6 +57,6 @@ public override IAsyncEnumerable<StreamingChatCompletionUpdate> CompleteStreamin
         options.StopSequences ??= [];
         options.StopSequences.Add(_eotToken);
 
-        return base.CompleteStreamingAsync(chatMessages, options, cancellationToken);
+        return base.GetStreamingResponseAsync(chatMessages, options, cancellationToken);
     }
 }