add deferred tool loading for kimi-k3

packages/ai/scripts/generate-models.ts

@@ -1675,6 +1675,11 @@ async function loadModelsDevData(): Promise<Model<any>[]> {
 				if (m.tool_call !== true) continue;
 
 				const isKimiK3 = modelId === "kimi-k3";
+				const compat = isKimiK3 ? { ...moonshotCompat } : moonshotCompat;
+				if (isKimiK3) {
+					compat.requiresReasoningContentOnAssistantMessages = true;
+					compat.deferredToolsMode = "kimi";
+				}
 				models.push({
 					id: modelId,
 					name: m.name || modelId,
@@ -1692,9 +1697,7 @@ async function loadModelsDevData(): Promise<Model<any>[]> {
 					},
 					contextWindow: m.limit?.context || 4096,
 					maxTokens: m.limit?.output || 4096,
-					compat: isKimiK3
+					compat,
-						? { ...moonshotCompat, requiresReasoningContentOnAssistantMessages: true }
-						: moonshotCompat,
 				});
 			}
 		}

packages/ai/src/providers/moonshotai-cn.models.ts

@@ -174,7 +174,7 @@ export const MOONSHOTAI_CN_MODELS = {
 		api: "openai-completions",
 		provider: "moonshotai-cn",
 		baseUrl: "https://api.moonshot.cn/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"thinkingFormat":"deepseek","requiresReasoningContentOnAssistantMessages":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"thinkingFormat":"deepseek","requiresReasoningContentOnAssistantMessages":true,"deferredToolsMode":"kimi"},
 		reasoning: true,
 		thinkingLevelMap: {"off":null,"minimal":null,"low":null,"medium":null,"high":null,"xhigh":null,"max":"max"},
 		input: ["text", "image"],

packages/ai/src/providers/moonshotai.models.ts

@@ -174,7 +174,7 @@ export const MOONSHOTAI_MODELS = {
 		api: "openai-completions",
 		provider: "moonshotai",
 		baseUrl: "https://api.moonshot.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"thinkingFormat":"deepseek","requiresReasoningContentOnAssistantMessages":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"thinkingFormat":"deepseek","requiresReasoningContentOnAssistantMessages":true,"deferredToolsMode":"kimi"},
 		reasoning: true,
 		thinkingLevelMap: {"off":null,"minimal":null,"low":null,"medium":null,"high":null,"xhigh":null,"max":"max"},
 		input: ["text", "image"],

packages/ai/src/providers/openrouter.models.ts

@@ -1923,6 +1923,24 @@ export const OPENROUTER_MODELS = {
 		contextWindow: 262144,
 		maxTokens: 262144,
 	} satisfies Model<"openai-completions">,
+	"moonshotai/kimi-k3": {
+		id: "moonshotai/kimi-k3",
+		name: "MoonshotAI: Kimi K3",
+		api: "openai-completions",
+		provider: "openrouter",
+		baseUrl: "https://openrouter.ai/api/v1",
+		compat: {"supportsDeveloperRole":false,"thinkingFormat":"openrouter"},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 3,
+			output: 15,
+			cacheRead: 0.3,
+			cacheWrite: 0,
+		},
+		contextWindow: 1048576,
+		maxTokens: 4096,
+	} satisfies Model<"openai-completions">,
 	"nex-agi/nex-n2-mini": {
 		id: "nex-agi/nex-n2-mini",
 		name: "Nex AGI: Nex-N2-Mini",
@@ -4688,9 +4706,9 @@ export const OPENROUTER_MODELS = {
 		thinkingLevelMap: {"xhigh":"xhigh"},
 		input: ["text"],
 		cost: {
-			input: 0.9366,
+			input: 0.9338,
-			output: 2.9436,
+			output: 2.9348,
-			cacheRead: 0.17394,
+			cacheRead: 0.17342,
 			cacheWrite: 0,
 		},
 		contextWindow: 1048576,
@@ -4832,13 +4850,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
-			input: 0.66,
+			input: 3,
-			output: 3.41,
+			output: 15,
-			cacheRead: 0.15,
+			cacheRead: 0.3,
 			cacheWrite: 0,
 		},
-		contextWindow: 262144,
+		contextWindow: 1048576,
-		maxTokens: 262144,
+		maxTokens: 4096,
 	} satisfies Model<"openai-completions">,
 	"~openai/gpt-latest": {
 		id: "~openai/gpt-latest",

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

@@ -1890,6 +1890,23 @@ export const VERCEL_AI_GATEWAY_MODELS = {
 		contextWindow: 262144,
 		maxTokens: 32768,
 	} satisfies Model<"anthropic-messages">,
+	"moonshotai/kimi-k3": {
+		id: "moonshotai/kimi-k3",
+		name: "Kimi K3",
+		api: "anthropic-messages",
+		provider: "vercel-ai-gateway",
+		baseUrl: "https://ai-gateway.vercel.sh",
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 3,
+			output: 15,
+			cacheRead: 0.3,
+			cacheWrite: 0,
+		},
+		contextWindow: 1000000,
+		maxTokens: 256000,
+	} satisfies Model<"anthropic-messages">,
 	"nvidia/nemotron-3-nano-30b-a3b": {
 		id: "nvidia/nemotron-3-nano-30b-a3b",
 		name: "Nemotron 3 Nano 30B A3B",