fix(aiplatform): update Gemini model from 2.0-flash to gemini-2.5-flash (#3342)

# fix(aiplatform): update Gemini models and rewrite image generation
sample

## Description
This PR updates the model references and APIs used across the .NET AI
Platform samples (`aiplatform/api/AIPlatform.Samples`) and integration
tests to align with current Vertex AI standards:

- **Gemini Models Update:** Upgrades text and multimodal model
references across all samples and tests from `gemini-2.0-flash-001` to
`gemini-2.5-flash`.
- **Image Sample Refactor:** Rewrites `GenerateImage.cs` to use the
modern Generative AI API (`GenerateContentAsync`) and the
`gemini-2.5-flash-image` model. This replaces the legacy `PredictAsync`
API pattern and JSON-encoded structures.

**Reason for Update:**
- `gemini-2.0-flash-001` reached end-of-life and was decommissioned on
June 1, 2026.
- The legacy `imagegeneration@006` model was also decommissioned,
causing `NotFound` RPC errors in CI
(`GenerateImageTest.TestGenerateImageAsync`). Migrating to
`gemini-2.5-flash-image` via the `GenerateContent` API ensures active
service availability and fixes the CI pipeline.

Fixes Internal: b/542706371

Author: angelcaamal

aiplatform/api/AIPlatform.Samples/AudioInputSummarization.cs

@@ -26,7 +26,7 @@ public async Task<string> SummarizeAudio(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001")
+        string model = "gemini-2.5-flash")
     {
         var predictionServiceClient = new PredictionServiceClientBuilder
         {

aiplatform/api/AIPlatform.Samples/AudioInputTranscription.cs

@@ -26,7 +26,7 @@ public async Task<string> TranscribeAudio(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001")
+        string model = "gemini-2.5-flash")
     {
 
         var predictionServiceClient = new PredictionServiceClientBuilder

aiplatform/api/AIPlatform.Samples/ContextCaching/CreateContextCache.cs

@@ -35,7 +35,7 @@ public async Task<CachedContentName> Create(string projectId)
             Parent = $"projects/{projectId}/locations/us-central1",
             CachedContent = new CachedContent
             {
-                Model = $"projects/{projectId}/locations/us-central1/publishers/google/models/gemini-2.0-flash-001",
+                Model = $"projects/{projectId}/locations/us-central1/publishers/google/models/gemini-2.5-flash",
                 SystemInstruction = new Content
                 {
                     Parts =

aiplatform/api/AIPlatform.Samples/ContextCaching/UseContextCache.cs

@@ -32,7 +32,7 @@ public async Task<string> Use(string projectId, CachedContentName name)
         var generateContentRequest = new GenerateContentRequest
         {
             CachedContentAsCachedContentName = name,
-            Model = $"projects/{projectId}/locations/us-central1/publishers/google/models/gemini-2.0-flash-001",
+            Model = $"projects/{projectId}/locations/us-central1/publishers/google/models/gemini-2.5-flash",
             Contents =
             {
                 new Content

aiplatform/api/AIPlatform.Samples/ControlledGeneration.cs

@@ -26,7 +26,7 @@ public async Task<string> GenerateContentWithResponseMimeType(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001")
+        string model = "gemini-2.5-flash")
     {
 
         var predictionServiceClient = new PredictionServiceClientBuilder
@@ -73,7 +73,7 @@ public async Task<string> GenerateContentWithResponseSchema(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001")
+        string model = "gemini-2.5-flash")
     {
 
         var predictionServiceClient = new PredictionServiceClientBuilder
@@ -130,7 +130,7 @@ public async Task<string> GenerateContentWithResponseSchema2(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001")
+        string model = "gemini-2.5-flash")
     {
 
         var predictionServiceClient = new PredictionServiceClientBuilder
@@ -194,7 +194,7 @@ public async Task<string> GenerateContentWithResponseSchema3(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001")
+        string model = "gemini-2.5-flash")
     {
 
         var predictionServiceClient = new PredictionServiceClientBuilder
@@ -270,7 +270,7 @@ public async Task<string> GenerateContentWithResponseSchema4(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001")
+        string model = "gemini-2.5-flash")
     {
 
         var predictionServiceClient = new PredictionServiceClientBuilder
@@ -362,7 +362,7 @@ public async Task<string> GenerateContentWithResponseSchema6(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001")
+        string model = "gemini-2.5-flash")
     {
 
         var predictionServiceClient = new PredictionServiceClientBuilder

aiplatform/api/AIPlatform.Samples/FunctionCalling.cs

@@ -28,7 +28,7 @@ public async Task<string> GenerateFunctionCall(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001")
+        string model = "gemini-2.5-flash")
     {
         var predictionServiceClient = new PredictionServiceClientBuilder
         {

aiplatform/api/AIPlatform.Samples/GeminiQuickstart.cs

@@ -27,7 +27,7 @@ public async Task<string> GenerateContent(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001"
+        string model = "gemini-2.5-flash"
     )
     {
         // Create client

aiplatform/api/AIPlatform.Samples/GenerateImage.cs

@@ -21,7 +21,6 @@
 using System.IO;
 using System.Linq;
 using System.Threading.Tasks;
-using Value = Google.Protobuf.WellKnownTypes.Value;
 
 public class GenerateImage
 {
@@ -33,35 +32,37 @@ public async Task<FileInfo> Generate(
             Endpoint = "us-central1-aiplatform.googleapis.com"
         }.Build();
 
-
         string prompt = "a dog reading a newspaper";
         string outputFileName = "dog_newspaper.png";
-        string model = "imagegeneration@006";
+        string location = "us-central1";
+        string model = "gemini-2.5-flash-image";
 
-        var predictRequest = new PredictRequest
+        var generateContentRequest = new GenerateContentRequest
         {
-            EndpointAsEndpointName = EndpointName.FromProjectLocationPublisherModel(projectId, "us-central1", "google", model),
-            Instances =
+            Model = $"projects/{projectId}/locations/{location}/publishers/google/models/{model}",
+            Contents =
             {
-                Value.ForStruct(new()
+                new Content
                 {
-                    Fields =
-                    {
-                        ["prompt"] = Value.ForString(prompt)
+                    Role = "USER",
+                    Parts = 
+                    { 
+                        new Part { Text = prompt } 
                     }
-                })
-            },
-            Parameters = Value.ForStruct(new()
-            {
-                Fields =
-                {
-                    ["sampleCount"] = Value.ForNumber(1)
                 }
-            })
+            }
         };
 
-        PredictResponse response = await predictionServiceClient.PredictAsync(predictRequest);
-        byte[] imageBytes = Convert.FromBase64String(response.Predictions.First().StructValue.Fields["bytesBase64Encoded"].StringValue);
+        GenerateContentResponse response = await predictionServiceClient.GenerateContentAsync(generateContentRequest);
+
+        var imagePart = response.Candidates.FirstOrDefault()?.Content?.Parts?.FirstOrDefault(p => p.InlineData != null);
+        
+        if (imagePart == null)
+        {
+            throw new Exception("No image data was returned by the model.");
+        }
+
+        byte[] imageBytes = imagePart.InlineData.Data.ToByteArray();
 
         File.WriteAllBytes(outputFileName, imageBytes);
         FileInfo fileInfo = new FileInfo(Path.GetFullPath(outputFileName));
@@ -71,4 +72,4 @@ public async Task<FileInfo> Generate(
     }
 }
 
-// [END generativeaionvertexai_imagen_generate_image]
+// [END generativeaionvertexai_imagen_generate_image]

aiplatform/api/AIPlatform.Samples/GetTokenCount.cs

@@ -27,7 +27,7 @@ public async Task<int> CountTokens(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001"
+        string model = "gemini-2.5-flash"
     )
     {
         var client = new LlmUtilityServiceClientBuilder

aiplatform/api/AIPlatform.Samples/GroundingVertexAiSearchSample.cs

@@ -26,7 +26,7 @@ public async Task<string> GenerateTextWithVertexAiSearch(
         string projectId = "your-project-id",
         string location = "us-central1",
         string publisher = "google",
-        string model = "gemini-2.0-flash-001",
+        string model = "gemini-2.5-flash",
         string dataStoreLocation = "global",
         string dataStoreId = "your-datastore-id")
     {