fix: update gemini model to 2.5-flash and increase MaxTokens

gemini-2.0-flash is no longer available to new API keys. Update all
integration tests to use gemini-2.5-flash and increase MaxTokens
since thinking models use tokens for internal reasoning.

Author: priyanshujain

agent/integration_test.go

@@ -62,7 +62,7 @@ func availableProviders(t *testing.T) []providerTestCase {
 		providers = append(providers, providerTestCase{
 			name:     "gemini",
 			provider: gemini.New(key),
-			model:    "gemini-2.0-flash",
+			model:    "gemini-2.5-flash",
 		})
 	}
 	if project := os.Getenv("GOOGLE_CLOUD_PROJECT"); project != "" {
@@ -74,7 +74,7 @@ func availableProviders(t *testing.T) []providerTestCase {
 		providers = append(providers, providerTestCase{
 			name:     "gemini-vertex",
 			provider: gemini.New("", gemini.WithVertexAI(project, region), gemini.WithTokenSource(provider.GcloudTokenSource(account))),
-			model:    "gemini-2.0-flash",
+			model:    "gemini-2.5-flash",
 		})
 	}
 
@@ -213,7 +213,7 @@ func TestIntegration_Streaming(t *testing.T) {
 			ch, err := tc.provider.StreamChat(ctx, provider.ChatRequest{
 				Model:     tc.model,
 				Messages:  []provider.Message{provider.NewTextMessage(provider.RoleUser, "Say 'test-stream-ok' and nothing else.")},
-				MaxTokens: 32,
+				MaxTokens: 256,
 			})
 			if err != nil {
 				t.Fatalf("StreamChat: %v", err)

internal/cli/setup_models.go

@@ -49,7 +49,7 @@ var llmProviders = []providerInfo{
 		models: []huh.Option[string]{
 			huh.NewOption("gemini-2.5-pro (most capable)", "gemini-2.5-pro"),
 			huh.NewOption("gemini-2.5-flash (fast, good balance)", "gemini-2.5-flash"),
-			huh.NewOption("gemini-2.0-flash (fastest, cheapest)", "gemini-2.0-flash"),
+			huh.NewOption("gemini-2.0-flash-lite (fastest, cheapest)", "gemini-2.0-flash-lite"),
 		},
 	},
 }

memory/integration_test.go

@@ -39,7 +39,7 @@ func availableProviders(t *testing.T) []providerTestCase {
 		providers = append(providers, providerTestCase{
 			name:     "gemini",
 			provider: gemini.New(key),
-			model:    "gemini-2.0-flash",
+			model:    "gemini-2.5-flash",
 		})
 	}
 	if project := os.Getenv("GOOGLE_CLOUD_PROJECT"); project != "" {
@@ -51,7 +51,7 @@ func availableProviders(t *testing.T) []providerTestCase {
 		providers = append(providers, providerTestCase{
 			name:     "gemini-vertex",
 			provider: gemini.New("", gemini.WithVertexAI(project, region), gemini.WithTokenSource(provider.GcloudTokenSource(account))),
-			model:    "gemini-2.0-flash",
+			model:    "gemini-2.5-flash",
 		})
 	}
 

provider/gemini/gemini_test.go

@@ -29,7 +29,7 @@ func TestChat_TextResponse(t *testing.T) {
 
 	p := New("test-key", WithBaseURL(server.URL))
 	resp, err := p.Chat(context.Background(), provider.ChatRequest{
-		Model:    "gemini-2.0-flash",
+		Model:    "gemini-2.5-flash",
 		Messages: []provider.Message{provider.NewTextMessage(provider.RoleUser, "Hello")},
 	})
 	if err != nil {
@@ -67,7 +67,7 @@ func TestChat_ToolUse(t *testing.T) {
 
 	p := New("test-key", WithBaseURL(server.URL))
 	resp, err := p.Chat(context.Background(), provider.ChatRequest{
-		Model:    "gemini-2.0-flash",
+		Model:    "gemini-2.5-flash",
 		Messages: []provider.Message{provider.NewTextMessage(provider.RoleUser, "run echo hi")},
 	})
 	if err != nil {
@@ -96,7 +96,7 @@ func TestChat_ErrorResponse(t *testing.T) {
 
 	p := New("bad-key", WithBaseURL(server.URL))
 	_, err := p.Chat(context.Background(), provider.ChatRequest{
-		Model:    "gemini-2.0-flash",
+		Model:    "gemini-2.5-flash",
 		Messages: []provider.Message{provider.NewTextMessage(provider.RoleUser, "Hi")},
 	})
 	if err == nil {
@@ -124,7 +124,7 @@ func TestStreamChat_TextDelta(t *testing.T) {
 
 	p := New("test-key", WithBaseURL(server.URL))
 	ch, err := p.StreamChat(context.Background(), provider.ChatRequest{
-		Model:    "gemini-2.0-flash",
+		Model:    "gemini-2.5-flash",
 		Messages: []provider.Message{provider.NewTextMessage(provider.RoleUser, "Hi")},
 	})
 	if err != nil {
@@ -150,9 +150,9 @@ func TestGeminiIntegration(t *testing.T) {
 
 	p := New(apiKey)
 	resp, err := p.Chat(context.Background(), provider.ChatRequest{
-		Model:     "gemini-2.0-flash",
+		Model:     "gemini-2.5-flash",
 		Messages:  []provider.Message{provider.NewTextMessage(provider.RoleUser, "Say 'hello' and nothing else.")},
-		MaxTokens: 32,
+		MaxTokens: 256,
 	})
 	if err != nil {
 		t.Fatalf("Chat: %v", err)
@@ -178,9 +178,9 @@ func TestGeminiVertexAIIntegration(t *testing.T) {
 
 	p := New("", WithVertexAI(project, region), WithTokenSource(provider.GcloudTokenSource(account)))
 	resp, err := p.Chat(context.Background(), provider.ChatRequest{
-		Model:     "gemini-2.0-flash",
+		Model:     "gemini-2.5-flash",
 		Messages:  []provider.Message{provider.NewTextMessage(provider.RoleUser, "Say 'hello' and nothing else.")},
-		MaxTokens: 32,
+		MaxTokens: 256,
 	})
 	if err != nil {
 		t.Fatalf("Chat: %v", err)