feat: update Gemini and Vertex AI models from preview to GA version (#5444)

- Remove outdated preview models (gemini-2.5-pro-preview-*) from type definitions
- Update to GA model (gemini-2.5-pro) across all providers
- Implement backward compatibility for existing users with preview models
- Update OpenRouter integration to remove preview model references
- Fix related tests to reflect model changes

Fixes #5444

packages/types/src/providers/gemini.ts

@@ -68,80 +68,6 @@ export const geminiModels = {
 		inputPrice: 0,
 		outputPrice: 0,
 	},
-	"gemini-2.5-pro-preview-03-25": {
-		maxTokens: 65_535,
-		contextWindow: 1_048_576,
-		supportsImages: true,
-		supportsPromptCache: true,
-		inputPrice: 2.5, // This is the pricing for prompts above 200k tokens.
-		outputPrice: 15,
-		cacheReadsPrice: 0.625,
-		cacheWritesPrice: 4.5,
-		tiers: [
-			{
-				contextWindow: 200_000,
-				inputPrice: 1.25,
-				outputPrice: 10,
-				cacheReadsPrice: 0.31,
-			},
-			{
-				contextWindow: Infinity,
-				inputPrice: 2.5,
-				outputPrice: 15,
-				cacheReadsPrice: 0.625,
-			},
-		],
-	},
-	"gemini-2.5-pro-preview-05-06": {
-		maxTokens: 65_535,
-		contextWindow: 1_048_576,
-		supportsImages: true,
-		supportsPromptCache: true,
-		inputPrice: 2.5, // This is the pricing for prompts above 200k tokens.
-		outputPrice: 15,
-		cacheReadsPrice: 0.625,
-		cacheWritesPrice: 4.5,
-		tiers: [
-			{
-				contextWindow: 200_000,
-				inputPrice: 1.25,
-				outputPrice: 10,
-				cacheReadsPrice: 0.31,
-			},
-			{
-				contextWindow: Infinity,
-				inputPrice: 2.5,
-				outputPrice: 15,
-				cacheReadsPrice: 0.625,
-			},
-		],
-	},
-	"gemini-2.5-pro-preview-06-05": {
-		maxTokens: 65_535,
-		contextWindow: 1_048_576,
-		supportsImages: true,
-		supportsPromptCache: true,
-		inputPrice: 2.5, // This is the pricing for prompts above 200k tokens.
-		outputPrice: 15,
-		cacheReadsPrice: 0.625,
-		cacheWritesPrice: 4.5,
-		maxThinkingTokens: 32_768,
-		supportsReasoningBudget: true,
-		tiers: [
-			{
-				contextWindow: 200_000,
-				inputPrice: 1.25,
-				outputPrice: 10,
-				cacheReadsPrice: 0.31,
-			},
-			{
-				contextWindow: Infinity,
-				inputPrice: 2.5,
-				outputPrice: 15,
-				cacheReadsPrice: 0.625,
-			},
-		],
-	},
 	"gemini-2.5-pro": {
 		maxTokens: 64_000,
 		contextWindow: 1_048_576,

packages/types/src/providers/openrouter.ts

@@ -78,7 +78,6 @@ export const OPEN_ROUTER_REASONING_BUDGET_MODELS = new Set([
 	"anthropic/claude-3.7-sonnet:beta",
 	"anthropic/claude-opus-4",
 	"anthropic/claude-sonnet-4",
-	"google/gemini-2.5-pro-preview",
 	"google/gemini-2.5-pro",
 	"google/gemini-2.5-flash-preview-05-20",
 	"google/gemini-2.5-flash",

packages/types/src/providers/vertex.ts

@@ -56,32 +56,6 @@ export const vertexModels = {
 		inputPrice: 0.15,
 		outputPrice: 0.6,
 	},
-	"gemini-2.5-pro-preview-03-25": {
-		maxTokens: 65_535,
-		contextWindow: 1_048_576,
-		supportsImages: true,
-		supportsPromptCache: true,
-		inputPrice: 2.5,
-		outputPrice: 15,
-	},
-	"gemini-2.5-pro-preview-05-06": {
-		maxTokens: 65_535,
-		contextWindow: 1_048_576,
-		supportsImages: true,
-		supportsPromptCache: true,
-		inputPrice: 2.5,
-		outputPrice: 15,
-	},
-	"gemini-2.5-pro-preview-06-05": {
-		maxTokens: 65_535,
-		contextWindow: 1_048_576,
-		supportsImages: true,
-		supportsPromptCache: true,
-		inputPrice: 2.5,
-		outputPrice: 15,
-		maxThinkingTokens: 32_768,
-		supportsReasoningBudget: true,
-	},
 	"gemini-2.5-pro": {
 		maxTokens: 64_000,
 		contextWindow: 1_048_576,

packages/types/src/providers/vscode-llm.ts

@@ -120,8 +120,8 @@ export const vscodeLlmModels = {
 		inputPrice: 0,
 		outputPrice: 0,
 		family: "gemini-2.5-pro",
-		version: "gemini-2.5-pro-preview-03-25",
-		name: "Gemini 2.5 Pro (Preview)",
+		version: "gemini-2.5-pro",
+		name: "Gemini 2.5 Pro",
 		supportsToolCalling: true,
 		maxInputTokens: 63830,
 	},

src/api/providers/fetchers/__tests__/openrouter.spec.ts

@@ -223,7 +223,6 @@ describe("OpenRouter API", () => {
 					contextWindow: 1048576,
 					supportsImages: true,
 					supportsPromptCache: true,
-					supportsReasoningBudget: true,
 					inputPrice: 1.25,
 					outputPrice: 10,
 					cacheWritesPrice: 1.625,
@@ -237,7 +236,6 @@ describe("OpenRouter API", () => {
 					contextWindow: 1048576,
 					supportsImages: true,
 					supportsPromptCache: true,
-					supportsReasoningBudget: true,
 					inputPrice: 1.25,
 					outputPrice: 10,
 					cacheWritesPrice: 1.625,

src/api/providers/gemini.ts

@@ -131,7 +131,15 @@ export class GeminiHandler extends BaseProvider implements SingleCompletionHandl
 
 	override getModel() {
 		const modelId = this.options.apiModelId
-		let id = modelId && modelId in geminiModels ? (modelId as GeminiModelId) : geminiDefaultModelId
+
+		// Handle backward compatibility for legacy preview model names
+		let mappedModelId = modelId
+		if (modelId && this.isLegacyPreviewModel(modelId)) {
+			mappedModelId = "gemini-2.5-pro"
+		}
+
+		let id =
+			mappedModelId && mappedModelId in geminiModels ? (mappedModelId as GeminiModelId) : geminiDefaultModelId
 		const info: ModelInfo = geminiModels[id]
 		const params = getModelParams({ format: "gemini", modelId: id, model: info, settings: this.options })
 
@@ -142,6 +150,15 @@ export class GeminiHandler extends BaseProvider implements SingleCompletionHandl
 		return { id: id.endsWith(":thinking") ? id.replace(":thinking", "") : id, info, ...params }
 	}
 
+	protected isLegacyPreviewModel(modelId: string): boolean {
+		const legacyPreviewModels = [
+			"gemini-2.5-pro-preview-03-25",
+			"gemini-2.5-pro-preview-05-06",
+			"gemini-2.5-pro-preview-06-05",
+		]
+		return legacyPreviewModels.includes(modelId)
+	}
+
 	async completePrompt(prompt: string): Promise<string> {
 		try {
 			const { id: model } = this.getModel()

src/api/providers/openrouter.ts

@@ -82,10 +82,13 @@ export class OpenRouterHandler extends BaseProvider implements SingleCompletionH
 		// other providers (including Gemini), so we need to explicitly disable
 		// i We should generalize this using the logic in `getModelParams`, but
 		// this is easier for now.
-		if (
-			(modelId === "google/gemini-2.5-pro-preview" || modelId === "google/gemini-2.5-pro") &&
-			typeof reasoning === "undefined"
-		) {
+		// Handle backward compatibility for legacy preview model names
+		let mappedModelId = modelId
+		if (this.isLegacyGeminiPreviewModel(modelId)) {
+			mappedModelId = "google/gemini-2.5-pro"
+		}
+
+		if (mappedModelId === "google/gemini-2.5-pro" && typeof reasoning === "undefined") {
 			reasoning = { exclude: true }
 		}
 
@@ -242,4 +245,8 @@ export class OpenRouterHandler extends BaseProvider implements SingleCompletionH
 		const completion = response as OpenAI.Chat.ChatCompletion
 		return completion.choices[0]?.message?.content || ""
 	}
+
+	private isLegacyGeminiPreviewModel(modelId: string): boolean {
+		return modelId === "google/gemini-2.5-pro-preview"
+	}
 }

src/api/providers/vertex.ts

@@ -14,7 +14,15 @@ export class VertexHandler extends GeminiHandler implements SingleCompletionHand
 
 	override getModel() {
 		const modelId = this.options.apiModelId
-		let id = modelId && modelId in vertexModels ? (modelId as VertexModelId) : vertexDefaultModelId
+
+		// Handle backward compatibility for legacy preview model names
+		let mappedModelId = modelId
+		if (modelId && this.isLegacyPreviewModel(modelId)) {
+			mappedModelId = "gemini-2.5-pro"
+		}
+
+		let id =
+			mappedModelId && mappedModelId in vertexModels ? (mappedModelId as VertexModelId) : vertexDefaultModelId
 		const info: ModelInfo = vertexModels[id]
 		const params = getModelParams({ format: "gemini", modelId: id, model: info, settings: this.options })