test(bedrock): add test for Opus 4.7 adaptive thinking support (#12288)

packages/types/src/providers/bedrock.ts

@@ -167,6 +167,30 @@ export const bedrockModels = {
 			},
 		],
 	},
+	"anthropic.claude-opus-4-7": {
+		maxTokens: 8192,
+		contextWindow: 200_000, // Default 200K, extendable to 1M with beta flag 'context-1m-2025-08-07'
+		supportsImages: true,
+		supportsPromptCache: true,
+		supportsReasoningBudget: true,
+		inputPrice: 5.0, // $5 per million input tokens (≤200K context)
+		outputPrice: 25.0, // $25 per million output tokens (≤200K context)
+		cacheWritesPrice: 6.25, // $6.25 per million tokens
+		cacheReadsPrice: 0.5, // $0.50 per million tokens
+		minTokensPerCachePoint: 1024,
+		maxCachePoints: 4,
+		cachableFields: ["system", "messages", "tools"],
+		// Tiered pricing for extended context (requires beta flag 'context-1m-2025-08-07')
+		tiers: [
+			{
+				contextWindow: 1_000_000, // 1M tokens with beta flag
+				inputPrice: 10.0, // $10 per million input tokens (>200K context)
+				outputPrice: 37.5, // $37.50 per million output tokens (>200K context)
+				cacheWritesPrice: 12.5, // $12.50 per million tokens (>200K context)
+				cacheReadsPrice: 1.0, // $1.00 per million tokens (>200K context)
+			},
+		],
+	},
 	"anthropic.claude-opus-4-5-20251101-v1:0": {
 		maxTokens: 8192,
 		contextWindow: 200_000,
@@ -525,6 +549,7 @@ export const BEDROCK_1M_CONTEXT_MODEL_IDS = [
 	"anthropic.claude-sonnet-4-5-20250929-v1:0",
 	"anthropic.claude-sonnet-4-6",
 	"anthropic.claude-opus-4-6-v1",
+	"anthropic.claude-opus-4-7",
 ] as const
 
 // Amazon Bedrock models that support Global Inference profiles
@@ -535,13 +560,15 @@ export const BEDROCK_1M_CONTEXT_MODEL_IDS = [
 // - Claude Haiku 4.5
 // - Claude Opus 4.5
 // - Claude Opus 4.6
+// - Claude Opus 4.7
 export const BEDROCK_GLOBAL_INFERENCE_MODEL_IDS = [
 	"anthropic.claude-sonnet-4-20250514-v1:0",
 	"anthropic.claude-sonnet-4-5-20250929-v1:0",
 	"anthropic.claude-sonnet-4-6",
 	"anthropic.claude-haiku-4-5-20251001-v1:0",
 	"anthropic.claude-opus-4-5-20251101-v1:0",
 	"anthropic.claude-opus-4-6-v1",
+	"anthropic.claude-opus-4-7",
 ] as const
 
 // Amazon Bedrock Service Tier types

src/api/providers/__tests__/bedrock-reasoning.spec.ts

@@ -282,6 +282,66 @@ describe("AwsBedrockHandler - Extended Thinking", () => {
 			expect(reasoningChunks[1].text).toBe(" about this problem.")
 		})
 
+		it("should use adaptive thinking for Opus 4.7 instead of enabled with budget_tokens", async () => {
+			handler = new AwsBedrockHandler({
+				apiProvider: "bedrock",
+				apiModelId: "anthropic.claude-opus-4-7",
+				awsRegion: "us-east-1",
+				enableReasoningEffort: true,
+				modelMaxTokens: 8192,
+				modelMaxThinkingTokens: 4096,
+			})
+
+			mockSend.mockResolvedValue({
+				stream: (async function* () {
+					yield { messageStart: { role: "assistant" } }
+					yield {
+						contentBlockStart: {
+							content_block: { type: "thinking", thinking: "Adaptive thinking..." },
+							contentBlockIndex: 0,
+						},
+					}
+					yield {
+						contentBlockDelta: {
+							delta: { type: "thinking_delta", thinking: " reasoning complete." },
+						},
+					}
+					yield {
+						contentBlockStart: {
+							start: { text: "Here is the answer." },
+							contentBlockIndex: 1,
+						},
+					}
+					yield { metadata: { usage: { inputTokens: 100, outputTokens: 50 } } }
+				})(),
+			})
+
+			const messages = [{ role: "user" as const, content: "Test message" }]
+			const stream = handler.createMessage("System prompt", messages)
+
+			const chunks = []
+			for await (const chunk of stream) {
+				chunks.push(chunk)
+			}
+
+			// Verify Opus 4.7 uses adaptive thinking (not enabled with budget_tokens)
+			expect(mockSend).toHaveBeenCalledTimes(1)
+			expect(capturedPayload).toBeDefined()
+			expect(capturedPayload.additionalModelRequestFields).toBeDefined()
+			expect(capturedPayload.additionalModelRequestFields.thinking).toEqual({
+				type: "adaptive",
+			})
+			expect(capturedPayload.additionalModelRequestFields.output_config).toEqual({
+				effort: "high",
+			})
+
+			// Verify reasoning chunks were yielded
+			const reasoningChunks = chunks.filter((c) => c.type === "reasoning")
+			expect(reasoningChunks).toHaveLength(2)
+			expect((reasoningChunks[0] as any).text).toBe("Adaptive thinking...")
+			expect((reasoningChunks[1] as any).text).toBe(" reasoning complete.")
+		})
+
 		it("should support API key authentication", async () => {
 			handler = new AwsBedrockHandler({
 				apiProvider: "bedrock",

src/api/providers/bedrock.ts

@@ -61,14 +61,26 @@ interface BedrockInferenceConfig {
 // Define interface for Bedrock additional model request fields
 // This includes thinking configuration, 1M context beta, and other model-specific parameters
 interface BedrockAdditionalModelFields {
-	thinking?: {
-		type: "enabled"
-		budget_tokens: number
+	thinking?:
+		| {
+				type: "enabled"
+				budget_tokens: number
+		  }
+		| {
+				type: "adaptive"
+		  }
+	output_config?: {
+		effort?: "low" | "medium" | "high"
 	}
 	anthropic_beta?: string[]
 	[key: string]: any // Add index signature to be compatible with DocumentType
 }
 
+// Models that only support thinking.type: "adaptive" (not "enabled" with budget_tokens)
+const BEDROCK_ADAPTIVE_THINKING_ONLY_MODEL_IDS = [
+	"anthropic.claude-opus-4-7",
+] as const
+
 // Define interface for Bedrock payload
 interface BedrockPayload {
 	modelId: BedrockModelId | string
@@ -392,12 +404,28 @@ export class AwsBedrockHandler extends BaseProvider implements SingleCompletionH
 
 		if ((isThinkingExplicitlyEnabled || isThinkingEnabledBySettings) && modelConfig.info.supportsReasoningBudget) {
 			thinkingEnabled = true
-			additionalModelRequestFields = {
-				thinking: {
-					type: "enabled",
-					budget_tokens: metadata?.thinking?.maxThinkingTokens || modelConfig.reasoningBudget || 4096,
-				},
+
+			// Opus 4.7 only supports thinking.type: "adaptive" with output_config.effort
+			// (NOT "enabled" with budget_tokens which returns a 400 error)
+			const baseId = this.parseBaseModelId(modelConfig.id)
+			if (BEDROCK_ADAPTIVE_THINKING_ONLY_MODEL_IDS.includes(baseId as any)) {
+				additionalModelRequestFields = {
+					thinking: {
+						type: "adaptive",
+					},
+					output_config: {
+						effort: "high",
+					},
+				}
+			} else {
+				additionalModelRequestFields = {
+					thinking: {
+						type: "enabled",
+						budget_tokens: metadata?.thinking?.maxThinkingTokens || modelConfig.reasoningBudget || 4096,
+					},
+				}
 			}
+
 			logger.info("Extended thinking enabled for Bedrock request", {
 				ctx: "bedrock",
 				modelId: modelConfig.id,