[Feat] Add Claude Sonnet 5.5 to model providers

.changeset/add-claude-sonnet-5-5-support.md

@@ -0,0 +1,5 @@
+---
+"zoo-code": minor
+---
+
+Add Claude Sonnet 5.5 (`claude-sonnet-5-5`) support across the Anthropic, Amazon Bedrock, Vertex AI, OpenRouter, Requesty, and Vercel AI Gateway providers. Sonnet 5.5 has a native 1M-token context window, 128K max output tokens, $2/$10 per-million-token input/output pricing, and keeps the Sonnet 5 adaptive-thinking contract (no sampling parameters, effort-based reasoning).

packages/types/src/providers/anthropic.ts

@@ -48,6 +48,24 @@ export const anthropicModels = {
 		description:
 			"Claude Sonnet 5 is the best combination of speed and intelligence, optimized for coding, tool use, and agentic workflows.",
 	},
+	"claude-sonnet-5-5": {
+		maxTokens: 128_000, // Overridden to 8k if `enableReasoningEffort` is false.
+		contextWindow: 1_000_000, // 1M context window native (no beta header required)
+		supportsImages: true,
+		supportsPromptCache: true,
+		inputPrice: 2.0, // $2 per million input tokens
+		outputPrice: 10.0, // $10 per million output tokens
+		cacheWritesPrice: 2.5, // $2.50 per million tokens (5m cache write)
+		cacheReadsPrice: 0.2, // $0.20 per million tokens
+		// Sonnet 5.5 keeps the Sonnet 5 adaptive-thinking / binary-toggle
+		// convention on the direct Anthropic provider path: manual budget_tokens
+		// and non-default sampling parameters return a 400.
+		supportsReasoningBudget: true,
+		supportsReasoningBinary: true,
+		supportsTemperature: false,
+		description:
+			"Claude Sonnet 5.5 is the best combination of speed and intelligence, optimized for coding, tool use, and agentic workflows.",
+	},
 	"claude-sonnet-4-5": {
 		maxTokens: 64_000, // Overridden to 8k if `enableReasoningEffort` is false.
 		contextWindow: 200_000, // Default 200K, extendable to 1M with beta flag 'context-1m-2025-08-07'

packages/types/src/providers/bedrock.ts

@@ -70,6 +70,27 @@ export const bedrockModels = {
 		description:
 			"Claude Sonnet 5 is the best combination of speed and intelligence, optimized for coding, tool use, and agentic workflows.",
 	},
+	"anthropic.claude-sonnet-5-5": {
+		// Undated model ID exactly as documented for Bedrock; no date-suffix
+		// variant has been published yet.
+		maxTokens: 128_000,
+		supportsMaxTokens: true,
+		contextWindow: 1_000_000, // 1M context window native (no beta header required)
+		supportsImages: true,
+		supportsPromptCache: true,
+		supportsReasoningBudget: true,
+		supportsReasoningBinary: true,
+		supportsTemperature: false,
+		inputPrice: 2.0, // $2 per million input tokens
+		outputPrice: 10.0, // $10 per million output tokens
+		cacheWritesPrice: 2.5, // $2.50 per million tokens (5m cache write)
+		cacheReadsPrice: 0.2, // $0.20 per million tokens
+		minTokensPerCachePoint: 1024,
+		maxCachePoints: 4,
+		cachableFields: ["system", "messages", "tools"],
+		description:
+			"Claude Sonnet 5.5 is the best combination of speed and intelligence, optimized for coding, tool use, and agentic workflows.",
+	},
 	"amazon.nova-pro-v1:0": {
 		maxTokens: 5000,
 		contextWindow: 300_000,
@@ -680,14 +701,19 @@ export const BEDROCK_1M_CONTEXT_MODEL_IDS = [
 // intentionally absent.
 // https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-anthropic-claude-sonnet-5.html
 // https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-anthropic-claude-opus-5.html
-export const BEDROCK_THINKING_DISABLE_MODEL_IDS = ["anthropic.claude-sonnet-5", "anthropic.claude-opus-5"] as const
+export const BEDROCK_THINKING_DISABLE_MODEL_IDS = [
+	"anthropic.claude-sonnet-5",
+	"anthropic.claude-sonnet-5-5",
+	"anthropic.claude-opus-5",
+] as const
 
 // Amazon Bedrock models that support Global Inference profiles
 // As of Nov 2025, AWS supports Global Inference for:
 // - Claude Sonnet 4
 // - Claude Sonnet 4.5
 // - Claude Sonnet 4.6
 // - Claude Sonnet 5
+// - Claude Sonnet 5.5
 // - Claude Haiku 4.5
 // - Claude Opus 4.5
 // - Claude Opus 4.6
@@ -700,6 +726,7 @@ export const BEDROCK_GLOBAL_INFERENCE_MODEL_IDS = [
 	"anthropic.claude-sonnet-4-5-20250929-v1:0",
 	"anthropic.claude-sonnet-4-6",
 	"anthropic.claude-sonnet-5",
+	"anthropic.claude-sonnet-5-5",
 	"anthropic.claude-haiku-4-5-20251001-v1:0",
 	"anthropic.claude-opus-4-5-20251101-v1:0",
 	"anthropic.claude-opus-4-6-v1",

packages/types/src/providers/openrouter.ts

@@ -40,6 +40,7 @@ export const OPEN_ROUTER_PROMPT_CACHING_MODELS = new Set([
 	"anthropic/claude-sonnet-4.5",
 	"anthropic/claude-sonnet-4.6",
 	"anthropic/claude-sonnet-5",
+	"anthropic/claude-sonnet-5-5",
 	"anthropic/claude-opus-4",
 	"anthropic/claude-opus-4.1",
 	"anthropic/claude-opus-4.5",
@@ -87,6 +88,7 @@ export const OPEN_ROUTER_REASONING_BUDGET_MODELS = new Set([
 	"anthropic/claude-sonnet-4.5",
 	"anthropic/claude-sonnet-4.6",
 	"anthropic/claude-sonnet-5",
+	"anthropic/claude-sonnet-5-5",
 	"anthropic/claude-haiku-4.5",
 	"google/gemini-2.5-pro-preview",
 	"google/gemini-2.5-pro",

packages/types/src/providers/vercel-ai-gateway.ts

@@ -20,6 +20,7 @@ export const VERCEL_AI_GATEWAY_PROMPT_CACHING_MODELS = new Set([
 	"anthropic/claude-sonnet-4",
 	"anthropic/claude-sonnet-4.6",
 	"anthropic/claude-sonnet-5",
+	"anthropic/claude-sonnet-5-5",
 	"openai/gpt-4.1",
 	"openai/gpt-4.1-mini",
 	"openai/gpt-4.1-nano",
@@ -69,6 +70,7 @@ export const VERCEL_AI_GATEWAY_VISION_AND_TOOLS_MODELS = new Set([
 	"anthropic/claude-sonnet-4.5",
 	"anthropic/claude-sonnet-4.6",
 	"anthropic/claude-sonnet-5",
+	"anthropic/claude-sonnet-5-5",
 	"google/gemini-1.5-flash",
 	"google/gemini-1.5-pro",
 	"google/gemini-2.0-flash",

packages/types/src/providers/vertex.ts

@@ -396,6 +396,21 @@ export const vertexModels = {
 		description:
 			"Claude Sonnet 5 is the best combination of speed and intelligence, optimized for coding, tool use, and agentic workflows.",
 	},
+	"claude-sonnet-5-5": {
+		maxTokens: 128_000, // 128K max output tokens per the Anthropic model card (same across all platforms).
+		contextWindow: 1_000_000, // 1M context window native (no beta header required)
+		supportsImages: true,
+		supportsPromptCache: true,
+		inputPrice: 2.0, // $2 per million input tokens
+		outputPrice: 10.0, // $10 per million output tokens
+		cacheWritesPrice: 2.5, // $2.50 per million tokens (5m cache write)
+		cacheReadsPrice: 0.2, // $0.20 per million tokens
+		supportsReasoningBudget: true,
+		supportsReasoningBinary: true,
+		supportsTemperature: false,
+		description:
+			"Claude Sonnet 5.5 is the best combination of speed and intelligence, optimized for coding, tool use, and agentic workflows.",
+	},
 	"claude-haiku-4-5@20251001": {
 		maxTokens: 8192,
 		contextWindow: 200_000,

src/api/providers/__tests__/anthropic-vertex.spec.ts

@@ -1027,6 +1027,25 @@ describe("VertexHandler", () => {
 			expect(model.info.supportsTemperature).toBe(false)
 		})
 
+		it("should return Claude Sonnet 5.5 model info", () => {
+			const handler = new AnthropicVertexHandler({
+				apiModelId: "claude-sonnet-5-5",
+				vertexProjectId: "test-project",
+				vertexRegion: "us-central1",
+			})
+
+			const model = handler.getModel()
+			expect(model.id).toBe("claude-sonnet-5-5")
+			expect(model.info.maxTokens).toBe(128_000)
+			expect(model.info.contextWindow).toBe(1_000_000)
+			expect(model.info.inputPrice).toBe(2.0)
+			expect(model.info.outputPrice).toBe(10.0)
+			expect(model.info.supportsReasoningBinary).toBe(true)
+			expect(model.info.supportsReasoningBudget).toBe(true)
+			expect(model.info.supportsPromptCache).toBe(true)
+			expect(model.info.supportsTemperature).toBe(false)
+		})
+
 		it("should return Claude Opus 5 model info", () => {
 			const handler = new AnthropicVertexHandler({
 				apiModelId: "claude-opus-5",
@@ -1390,6 +1409,41 @@ describe("VertexHandler", () => {
 			expect(request.temperature).toBeUndefined()
 		})
 
+		it("should use adaptive thinking for Claude Sonnet 5.5", async () => {
+			const sonnetHandler = new AnthropicVertexHandler({
+				apiModelId: "claude-sonnet-5-5",
+				vertexProjectId: "test-project",
+				vertexRegion: "us-central1",
+				enableReasoningEffort: true,
+			})
+
+			const mockCreate = vitest
+				.fn()
+				.mockImplementation(async () =>
+					asyncStreamFrom([
+						{ type: "message_start", message: { usage: { input_tokens: 10, output_tokens: 5 } } },
+					]),
+				)
+			// The SDK client's overloaded `create` signature can't be assigned a
+			// vitest mock directly, so a structural double assertion is required.
+			;(sonnetHandler["client"].messages as unknown as { create: typeof mockCreate }).create = mockCreate
+
+			await sonnetHandler
+				.createMessage("You are a helpful assistant", [{ role: "user", content: "Hello" }])
+				.next()
+
+			expect(mockCreate).toHaveBeenCalledWith(
+				expect.objectContaining({
+					thinking: { type: "adaptive" },
+				}),
+				undefined,
+			)
+
+			const request = mockCreate.mock.calls[0][0]
+			expect(request.thinking).not.toHaveProperty("budget_tokens")
+			expect(request.temperature).toBeUndefined()
+		})
+
 		it("should use adaptive thinking for Claude Opus 5", async () => {
 			const opusHandler = new AnthropicVertexHandler({
 				apiModelId: "claude-opus-5",

src/api/providers/__tests__/anthropic.spec.ts

@@ -456,6 +456,31 @@ describe("AnthropicHandler", () => {
 			expect(requestOptions?.headers?.["anthropic-beta"]).toContain("prompt-caching-2024-07-31")
 		})
 
+		it("should use adaptive thinking for Claude Sonnet 5.5 when reasoning is enabled", async () => {
+			const sonnetHandler = new AnthropicHandler({
+				apiKey: "test-api-key",
+				apiModelId: "claude-sonnet-5-5",
+				enableReasoningEffort: true,
+				modelMaxTokens: 32768,
+			})
+
+			const stream = sonnetHandler.createMessage(systemPrompt, [
+				{
+					role: "user",
+					content: [{ type: "text" as const, text: "Hello" }],
+				},
+			])
+
+			await collectStream(stream)
+
+			const requestBody = mockCreate.mock.calls[mockCreate.mock.calls.length - 1]?.[0]
+			const requestOptions = mockCreate.mock.calls[mockCreate.mock.calls.length - 1]?.[1]
+			expect(requestBody?.thinking).toEqual({ type: "adaptive" })
+			expect(requestBody?.temperature).toBeUndefined()
+			expect(requestBody?.max_tokens).toBe(32768)
+			expect(requestOptions?.headers?.["anthropic-beta"]).toContain("prompt-caching-2024-07-31")
+		})
+
 		it("should use adaptive thinking for Claude Opus 5 when reasoning is enabled", async () => {
 			const opusHandler = new AnthropicHandler({
 				apiKey: "test-api-key",
@@ -725,6 +750,27 @@ describe("AnthropicHandler", () => {
 			expect(model.reasoningBudget).toBeUndefined()
 		})
 
+		it("should handle Claude Sonnet 5.5 model correctly", () => {
+			const handler = new AnthropicHandler({
+				apiKey: "test-api-key",
+				apiModelId: "claude-sonnet-5-5",
+			})
+			const model = handler.getModel()
+			expect(model.id).toBe("claude-sonnet-5-5")
+			expect(model.info.maxTokens).toBe(128000)
+			expect(model.info.contextWindow).toBe(1000000)
+			expect(model.info.inputPrice).toBe(2.0)
+			expect(model.info.outputPrice).toBe(10.0)
+			expect(model.info.cacheWritesPrice).toBe(2.5)
+			expect(model.info.cacheReadsPrice).toBe(0.2)
+			expect(model.maxTokens).toBe(8192)
+			expect(model.info.supportsReasoningBinary).toBe(true)
+			expect(model.info.supportsReasoningBudget).toBe(true)
+			expect(model.info.supportsPromptCache).toBe(true)
+			expect(model.info.supportsTemperature).toBe(false)
+			expect(model.reasoningBudget).toBeUndefined()
+		})
+
 		it("should handle Claude Opus 5 model correctly", () => {
 			const handler = new AnthropicHandler({
 				apiKey: "test-api-key",