Merge remote-tracking branch 'origin/main' into add-kimi-deferred-tools

packages/ai/src/api/openai-responses.ts

@@ -282,6 +282,7 @@ function buildParams(model: Model<"openai-responses">, context: Context, options
 				effort: (model.thinkingLevelMap?.off ?? "none") as NonNullable<typeof params.reasoning>["effort"],
 			};
 		}
+		if (model.provider === "xai") params.include = ["reasoning.encrypted_content"];
 	}
 
 	return params;

packages/ai/src/auth/oauth/load.ts

@@ -11,18 +11,46 @@ const importOAuthModule = (specifier: string): Promise<unknown> => {
 	return import(runtimeSpecifier);
 };
 
-export const loadAnthropicOAuth = async (): Promise<OAuthAuth> =>
-	((await importOAuthModule("./anthropic.ts")) as { anthropicOAuth: OAuthAuth }).anthropicOAuth;
+type OAuthFlowLoaders = {
+	anthropic: () => OAuthAuth | Promise<OAuthAuth>;
+	openaiCodex: () => OAuthAuth | Promise<OAuthAuth>;
+	githubCopilot: () => OAuthAuth | Promise<OAuthAuth>;
+	xai: () => OAuthAuth | Promise<OAuthAuth>;
+	radius: (options: { name: string; gateway: string }) => OAuthAuth | Promise<OAuthAuth>;
+};
 
-export const loadOpenAICodexOAuth = async (): Promise<OAuthAuth> =>
-	((await importOAuthModule("./openai-codex.ts")) as { openaiCodexOAuth: OAuthAuth }).openaiCodexOAuth;
+let bundledLoaders: OAuthFlowLoaders | undefined;
 
-export const loadGitHubCopilotOAuth = async (): Promise<OAuthAuth> =>
-	((await importOAuthModule("./github-copilot.ts")) as { githubCopilotOAuth: OAuthAuth }).githubCopilotOAuth;
+/** Registers statically bundled OAuth flows for standalone Bun binaries. */
+export function registerBundledOAuthFlowLoaders(loaders: OAuthFlowLoaders): void {
+	bundledLoaders = loaders;
+}
 
-export const loadRadiusOAuth = async (options: { name: string; gateway: string }): Promise<OAuthAuth> =>
-	(
+export const loadAnthropicOAuth = async (): Promise<OAuthAuth> => {
+	if (bundledLoaders) return bundledLoaders.anthropic();
+	return ((await importOAuthModule("./anthropic.ts")) as { anthropicOAuth: OAuthAuth }).anthropicOAuth;
+};
+
+export const loadOpenAICodexOAuth = async (): Promise<OAuthAuth> => {
+	if (bundledLoaders) return bundledLoaders.openaiCodex();
+	return ((await importOAuthModule("./openai-codex.ts")) as { openaiCodexOAuth: OAuthAuth }).openaiCodexOAuth;
+};
+
+export const loadGitHubCopilotOAuth = async (): Promise<OAuthAuth> => {
+	if (bundledLoaders) return bundledLoaders.githubCopilot();
+	return ((await importOAuthModule("./github-copilot.ts")) as { githubCopilotOAuth: OAuthAuth }).githubCopilotOAuth;
+};
+
+export const loadXaiOAuth = async (): Promise<OAuthAuth> => {
+	if (bundledLoaders) return bundledLoaders.xai();
+	return ((await importOAuthModule("./xai.ts")) as { xaiOAuth: OAuthAuth }).xaiOAuth;
+};
+
+export const loadRadiusOAuth = async (options: { name: string; gateway: string }): Promise<OAuthAuth> => {
+	if (bundledLoaders) return bundledLoaders.radius(options);
+	return (
 		(await importOAuthModule("./radius.ts")) as {
 			createRadiusOAuth: (input: { name: string; gateway: string }) => OAuthAuth;
 		}
 	).createRadiusOAuth(options);
+};

packages/ai/src/auth/oauth/xai.ts

@@ -0,0 +1,231 @@
+/**
+ * xAI OAuth device-code flow.
+ */
+
+import type { AuthInteraction, OAuthAuth, OAuthCredential } from "../types.ts";
+import { pollOAuthDeviceCodeFlow } from "./device-code.ts";
+
+const XAI_CLIENT_ID = "b1a00492-073a-47ea-816f-4c329264a828";
+const XAI_SCOPE = "openid profile email offline_access grok-cli:access api:access";
+const XAI_DEVICE_CODE_URL = "https://auth.x.ai/oauth2/device/code";
+const XAI_TOKEN_URL = "https://auth.x.ai/oauth2/token";
+// Refresh slightly before the reported expiry to avoid using a token that dies mid-request.
+const REFRESH_SKEW_MS = 5 * 60 * 1000;
+const DEFAULT_TOKEN_LIFETIME_SECONDS = 3600;
+
+type JsonObject = Record<string, unknown>;
+
+type OAuthHttpResponse = {
+	ok: boolean;
+	status: number;
+	body: JsonObject;
+};
+
+type XaiDeviceCode = {
+	deviceCode: string;
+	userCode: string;
+	verificationUri: string;
+	intervalSeconds?: number;
+	expiresInSeconds: number;
+};
+
+function requiredString(body: JsonObject, field: string): string {
+	const value = body[field];
+	if (typeof value !== "string" || value.length === 0) {
+		throw new Error(`Invalid xAI OAuth response field: ${field}`);
+	}
+	return value;
+}
+
+function positiveNumber(body: JsonObject, field: string): number {
+	const value = body[field];
+	if (typeof value !== "number" || !Number.isFinite(value) || value <= 0) {
+		throw new Error(`Invalid xAI OAuth response field: ${field}`);
+	}
+	return value;
+}
+
+// The verification URI is opened in the user's browser; force it to be an https URL
+// so a malicious response cannot make `open` launch something else.
+function validateVerificationUri(raw: string): string {
+	let url: URL;
+	try {
+		url = new URL(raw);
+	} catch {
+		throw new Error("Untrusted verification URI in xAI OAuth response");
+	}
+	if (url.protocol !== "https:") {
+		throw new Error("Untrusted verification URI in xAI OAuth response");
+	}
+	return url.href;
+}
+
+async function postForm(url: string, fields: Record<string, string>, signal?: AbortSignal): Promise<OAuthHttpResponse> {
+	let response: Response;
+	try {
+		response = await fetch(url, {
+			method: "POST",
+			headers: {
+				Accept: "application/json",
+				"Content-Type": "application/x-www-form-urlencoded",
+			},
+			body: new URLSearchParams(fields),
+			signal,
+		});
+	} catch (error) {
+		if (signal?.aborted) {
+			throw new Error("Login cancelled");
+		}
+		throw error;
+	}
+
+	let body: JsonObject;
+	try {
+		const parsed = (await response.json()) as unknown;
+		body = parsed && typeof parsed === "object" && !Array.isArray(parsed) ? (parsed as JsonObject) : {};
+	} catch {
+		if (signal?.aborted) {
+			throw new Error("Login cancelled");
+		}
+		throw new Error(`xAI OAuth returned invalid JSON (HTTP ${response.status})`);
+	}
+	return {
+		ok: response.ok,
+		status: response.status,
+		body,
+	};
+}
+
+function requestFailure(action: string, response: OAuthHttpResponse): Error {
+	const error = typeof response.body.error === "string" ? response.body.error : undefined;
+	const description =
+		typeof response.body.error_description === "string" ? response.body.error_description : undefined;
+	const detail = [error, description].filter(Boolean).join(": ");
+	return new Error(`xAI OAuth ${action} failed (HTTP ${response.status})${detail ? `: ${detail}` : ""}`);
+}
+
+function parseDeviceCode(body: JsonObject): XaiDeviceCode {
+	// RFC 8628 allows interval 0 (no minimum wait); fall back to the poller's
+	// default instead of failing on non-positive or malformed values.
+	const interval = body.interval;
+	const intervalSeconds =
+		typeof interval === "number" && Number.isFinite(interval) && interval > 0 ? interval : undefined;
+	return {
+		deviceCode: requiredString(body, "device_code"),
+		userCode: requiredString(body, "user_code"),
+		verificationUri: validateVerificationUri(requiredString(body, "verification_uri")),
+		intervalSeconds,
+		expiresInSeconds: positiveNumber(body, "expires_in"),
+	};
+}
+
+function credentialsFromTokenResponse(body: JsonObject, previousRefreshToken?: string): OAuthCredential {
+	const access = requiredString(body, "access_token");
+	// xAI may omit refresh_token on refresh when the token is not rotated.
+	const refresh =
+		body.refresh_token === undefined && previousRefreshToken
+			? previousRefreshToken
+			: requiredString(body, "refresh_token");
+	const expiresInSeconds =
+		body.expires_in === undefined ? DEFAULT_TOKEN_LIFETIME_SECONDS : positiveNumber(body, "expires_in");
+	return {
+		type: "oauth",
+		access,
+		refresh,
+		expires: Date.now() + expiresInSeconds * 1000 - REFRESH_SKEW_MS,
+	};
+}
+
+async function requestDeviceCode(signal?: AbortSignal): Promise<XaiDeviceCode> {
+	const response = await postForm(
+		XAI_DEVICE_CODE_URL,
+		{
+			client_id: XAI_CLIENT_ID,
+			scope: XAI_SCOPE,
+			referrer: "pi",
+		},
+		signal,
+	);
+	if (!response.ok) {
+		throw requestFailure("device authorization", response);
+	}
+	return parseDeviceCode(response.body);
+}
+
+async function pollForTokens(device: XaiDeviceCode, signal?: AbortSignal): Promise<OAuthCredential> {
+	return pollOAuthDeviceCodeFlow<OAuthCredential>({
+		intervalSeconds: device.intervalSeconds,
+		expiresInSeconds: device.expiresInSeconds,
+		waitBeforeFirstPoll: true,
+		signal,
+		poll: async () => {
+			const response = await postForm(
+				XAI_TOKEN_URL,
+				{
+					grant_type: "urn:ietf:params:oauth:grant-type:device_code",
+					client_id: XAI_CLIENT_ID,
+					device_code: device.deviceCode,
+				},
+				signal,
+			);
+
+			if (response.ok) {
+				return { status: "complete", value: credentialsFromTokenResponse(response.body) };
+			}
+
+			const error = response.body.error;
+			if (error === "authorization_pending") {
+				return { status: "pending" };
+			}
+			if (error === "slow_down") {
+				const interval = response.body.interval;
+				return { status: "slow_down", intervalSeconds: typeof interval === "number" ? interval : undefined };
+			}
+			if (error === "access_denied" || error === "authorization_denied") {
+				return { status: "failed", message: "xAI device authorization was denied" };
+			}
+			if (error === "expired_token") {
+				return { status: "failed", message: "xAI device code expired" };
+			}
+			return { status: "failed", message: requestFailure("device token polling", response).message };
+		},
+	});
+}
+
+async function loginXai(interaction: AuthInteraction): Promise<OAuthCredential> {
+	const device = await requestDeviceCode(interaction.signal);
+	interaction.notify({
+		type: "device_code",
+		userCode: device.userCode,
+		verificationUri: device.verificationUri,
+		intervalSeconds: device.intervalSeconds,
+		expiresInSeconds: device.expiresInSeconds,
+	});
+	return pollForTokens(device, interaction.signal);
+}
+
+async function refreshXaiToken(refreshToken: string, signal?: AbortSignal): Promise<OAuthCredential> {
+	const response = await postForm(
+		XAI_TOKEN_URL,
+		{
+			grant_type: "refresh_token",
+			client_id: XAI_CLIENT_ID,
+			refresh_token: refreshToken,
+		},
+		signal,
+	);
+	if (!response.ok) {
+		throw requestFailure("token refresh", response);
+	}
+	return credentialsFromTokenResponse(response.body, refreshToken);
+}
+
+export const xaiOAuth: OAuthAuth = {
+	name: "xAI (Grok/X subscription)",
+	login: loginXai,
+	refresh: (credential, signal) => refreshXaiToken(credential.refresh, signal),
+
+	async toAuth(credential) {
+		return { apiKey: credential.access };
+	},
+};

packages/ai/src/bun-oauth.ts

@@ -0,0 +1,17 @@
+import { anthropicOAuth } from "./auth/oauth/anthropic.ts";
+import { githubCopilotOAuth } from "./auth/oauth/github-copilot.ts";
+import { registerBundledOAuthFlowLoaders } from "./auth/oauth/load.ts";
+import { openaiCodexOAuth } from "./auth/oauth/openai-codex.ts";
+import { createRadiusOAuth } from "./auth/oauth/radius.ts";
+import { xaiOAuth } from "./auth/oauth/xai.ts";
+
+/** Register OAuth flows statically embedded in the standalone Bun binary. */
+export function registerBunOAuthFlows(): void {
+	registerBundledOAuthFlowLoaders({
+		anthropic: () => anthropicOAuth,
+		openaiCodex: () => openaiCodexOAuth,
+		githubCopilot: () => githubCopilotOAuth,
+		xai: () => xaiOAuth,
+		radius: createRadiusOAuth,
+	});
+}

packages/ai/src/models.ts

@@ -38,11 +38,15 @@ export interface RefreshModelsContext {
 	store: ProviderModelsStore;
 	/** False during offline/cache-only initialization. */
 	allowNetwork: boolean;
+	/** Bypass provider freshness checks and fetch immediately when network access is allowed. */
+	force?: boolean;
 	signal?: AbortSignal;
 }
 
 export interface ModelsRefreshOptions {
 	allowNetwork?: boolean;
+	/** Bypass provider freshness checks and fetch immediately when network access is allowed. */
+	force?: boolean;
 	signal?: AbortSignal;
 }
 
@@ -290,7 +294,13 @@ class ModelsImpl implements MutableModels {
 					stored = await this.readCredential(provider.id);
 					const credential = await this.resolveRefreshCredential(provider, stored, allowNetwork, options.signal);
 					if (!credential) return;
-					await provider.refreshModels({ credential, store, allowNetwork, signal: options.signal });
+					await provider.refreshModels({
+						credential,
+						store,
+						allowNetwork,
+						force: options.force,
+						signal: options.signal,
+					});
 				} catch (error) {
 					if (!options.signal?.aborted) {
 						errors.set(

packages/ai/src/providers/kimi-coding.models.ts

@@ -22,6 +22,24 @@ export const KIMI_CODING_MODELS = {
 		contextWindow: 262144,
 		maxTokens: 32768,
 	} satisfies Model<"anthropic-messages">,
+	"k3": {
+		id: "k3",
+		name: "Kimi K3",
+		api: "anthropic-messages",
+		provider: "kimi-coding",
+		baseUrl: "https://api.kimi.com/coding",
+		headers: {"User-Agent":"KimiCLI/1.5"},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0,
+			output: 0,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 1048576,
+		maxTokens: 131072,
+	} satisfies Model<"anthropic-messages">,
 	"kimi-for-coding": {
 		id: "kimi-for-coding",
 		name: "Kimi For Coding",
@@ -40,6 +58,24 @@ export const KIMI_CODING_MODELS = {
 		contextWindow: 262144,
 		maxTokens: 32768,
 	} satisfies Model<"anthropic-messages">,
+	"kimi-for-coding-highspeed": {
+		id: "kimi-for-coding-highspeed",
+		name: "Kimi For Coding HighSpeed",
+		api: "anthropic-messages",
+		provider: "kimi-coding",
+		baseUrl: "https://api.kimi.com/coding",
+		headers: {"User-Agent":"KimiCLI/1.5"},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0,
+			output: 0,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 32768,
+	} satisfies Model<"anthropic-messages">,
 	"kimi-k2-thinking": {
 		id: "kimi-k2-thinking",
 		name: "Kimi K2 Thinking",

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

@@ -168,4 +168,23 @@ export const MOONSHOTAI_CN_MODELS = {
 		contextWindow: 262144,
 		maxTokens: 262144,
 	} satisfies Model<"openai-completions">,
+	"kimi-k3": {
+		id: "kimi-k3",
+		name: "Kimi K3",
+		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},
+		reasoning: true,
+		thinkingLevelMap: {"off":null,"minimal":null,"low":null,"medium":null,"high":null,"xhigh":null,"max":"max"},
+		input: ["text", "image"],
+		cost: {
+			input: 0,
+			output: 0,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 1048576,
+		maxTokens: 131072,
+	} satisfies Model<"openai-completions">,
 } as const;

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

@@ -168,4 +168,23 @@ export const MOONSHOTAI_MODELS = {
 		contextWindow: 262144,
 		maxTokens: 262144,
 	} satisfies Model<"openai-completions">,
+	"kimi-k3": {
+		id: "kimi-k3",
+		name: "Kimi K3",
+		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},
+		reasoning: true,
+		thinkingLevelMap: {"off":null,"minimal":null,"low":null,"medium":null,"high":null,"xhigh":null,"max":"max"},
+		input: ["text", "image"],
+		cost: {
+			input: 0,
+			output: 0,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 1048576,
+		maxTokens: 131072,
+	} satisfies Model<"openai-completions">,
 } as const;

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

@@ -385,7 +385,7 @@ export const OPENROUTER_MODELS = {
 			cacheRead: 0.3,
 			cacheWrite: 3.75,
 		},
-		contextWindow: 1000000,
+		contextWindow: 200000,
 		maxTokens: 64000,
 	} satisfies Model<"openai-completions">,
 	"anthropic/claude-sonnet-4.5": {
@@ -652,13 +652,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: false,
 		input: ["text"],
 		cost: {
-			input: 0.24,
-			output: 0.9,
+			input: 0.27,
+			output: 1.12,
 			cacheRead: 0.135,
 			cacheWrite: 0,
 		},
 		contextWindow: 163840,
-		maxTokens: 16384,
+		maxTokens: 65536,
 	} satisfies Model<"openai-completions">,
 	"deepseek/deepseek-chat-v3.1": {
 		id: "deepseek/deepseek-chat-v3.1",
@@ -725,11 +725,11 @@ export const OPENROUTER_MODELS = {
 		input: ["text"],
 		cost: {
 			input: 0.27,
-			output: 0.95,
-			cacheRead: 0.13,
+			output: 1,
+			cacheRead: 0.135,
 			cacheWrite: 0,
 		},
-		contextWindow: 163840,
+		contextWindow: 131072,
 		maxTokens: 32768,
 	} satisfies Model<"openai-completions">,
 	"deepseek/deepseek-v3.2": {
@@ -742,13 +742,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 0.2145,
-			output: 0.32175,
-			cacheRead: 0.02145,
+			input: 0.269,
+			output: 0.4,
+			cacheRead: 0.1345,
 			cacheWrite: 0,
 		},
-		contextWindow: 128000,
-		maxTokens: 64000,
+		contextWindow: 163840,
+		maxTokens: 65536,
 	} satisfies Model<"openai-completions">,
 	"deepseek/deepseek-v3.2-exp": {
 		id: "deepseek/deepseek-v3.2-exp",
@@ -779,13 +779,13 @@ export const OPENROUTER_MODELS = {
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null,"high":"high","max":null,"xhigh":"xhigh"},
 		input: ["text"],
 		cost: {
-			input: 0.09,
-			output: 0.18,
-			cacheRead: 0.018,
+			input: 0.098,
+			output: 0.196,
+			cacheRead: 0.02,
 			cacheWrite: 0,
 		},
-		contextWindow: 1048576,
-		maxTokens: 65536,
+		contextWindow: 1048575,
+		maxTokens: 4096,
 	} satisfies Model<"openai-completions">,
 	"deepseek/deepseek-v4-pro": {
 		id: "deepseek/deepseek-v4-pro",
@@ -1051,12 +1051,12 @@ export const OPENROUTER_MODELS = {
 		input: ["text", "image"],
 		cost: {
 			input: 0.08,
-			output: 0.16,
-			cacheRead: 0,
+			output: 0.45,
+			cacheRead: 0.04,
 			cacheWrite: 0,
 		},
 		contextWindow: 131072,
-		maxTokens: 16384,
+		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
 	"google/gemma-4-26b-a4b-it": {
 		id: "google/gemma-4-26b-a4b-it",
@@ -1068,13 +1068,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
-			input: 0.06,
-			output: 0.33,
+			input: 0.1,
+			output: 0.3,
 			cacheRead: 0,
 			cacheWrite: 0,
 		},
-		contextWindow: 262144,
-		maxTokens: 4096,
+		contextWindow: 256000,
+		maxTokens: 256000,
 	} satisfies Model<"openai-completions">,
 	"google/gemma-4-26b-a4b-it:free": {
 		id: "google/gemma-4-26b-a4b-it:free",
@@ -1104,13 +1104,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
-			input: 0.06,
-			output: 0.35,
-			cacheRead: 0,
+			input: 0.22,
+			output: 0.55,
+			cacheRead: 0.12,
 			cacheWrite: 0,
 		},
 		contextWindow: 262144,
-		maxTokens: 8192,
+		maxTokens: 262144,
 	} satisfies Model<"openai-completions">,
 	"google/gemma-4-31b-it:free": {
 		id: "google/gemma-4-31b-it:free",
@@ -1303,13 +1303,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: false,
 		input: ["text"],
 		cost: {
-			input: 0.02,
-			output: 0.03,
-			cacheRead: 0,
+			input: 0.05,
+			output: 0.08,
+			cacheRead: 0.025,
 			cacheWrite: 0,
 		},
 		contextWindow: 131072,
-		maxTokens: 16384,
+		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
 	"meta-llama/llama-3.3-70b-instruct": {
 		id: "meta-llama/llama-3.3-70b-instruct",
@@ -1321,13 +1321,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: false,
 		input: ["text"],
 		cost: {
-			input: 0.1,
-			output: 0.32,
+			input: 0.13,
+			output: 0.4,
 			cacheRead: 0,
 			cacheWrite: 0,
 		},
 		contextWindow: 131072,
-		maxTokens: 16384,
+		maxTokens: 128000,
 	} satisfies Model<"openai-completions">,
 	"meta-llama/llama-3.3-70b-instruct:free": {
 		id: "meta-llama/llama-3.3-70b-instruct:free",
@@ -1383,6 +1383,24 @@ export const OPENROUTER_MODELS = {
 		contextWindow: 327680,
 		maxTokens: 16384,
 	} satisfies Model<"openai-completions">,
+	"meta/muse-spark-1.1": {
+		id: "meta/muse-spark-1.1",
+		name: "Meta: Muse Spark 1.1",
+		api: "openai-completions",
+		provider: "openrouter",
+		baseUrl: "https://openrouter.ai/api/v1",
+		compat: {"supportsDeveloperRole":false,"thinkingFormat":"openrouter"},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 1.25,
+			output: 4.25,
+			cacheRead: 0.15,
+			cacheWrite: 0,
+		},
+		contextWindow: 1048576,
+		maxTokens: 4096,
+	} satisfies Model<"openai-completions">,
 	"minimax/minimax-m1": {
 		id: "minimax/minimax-m1",
 		name: "MiniMax: MiniMax M1",
@@ -1393,7 +1411,7 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 0.4,
+			input: 0.55,
 			output: 2.2,
 			cacheRead: 0,
 			cacheWrite: 0,
@@ -1465,13 +1483,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 0.24,
-			output: 0.96,
-			cacheRead: 0,
+			input: 0.3,
+			output: 1.2,
+			cacheRead: 0.06,
 			cacheWrite: 0,
 		},
-		contextWindow: 196608,
-		maxTokens: 196608,
+		contextWindow: 204800,
+		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
 	"minimax/minimax-m3": {
 		id: "minimax/minimax-m3",
@@ -1488,8 +1506,8 @@ export const OPENROUTER_MODELS = {
 			cacheRead: 0.06,
 			cacheWrite: 0,
 		},
-		contextWindow: 1000000,
-		maxTokens: 131072,
+		contextWindow: 524288,
+		maxTokens: 512000,
 	} satisfies Model<"openai-completions">,
 	"mistralai/codestral-2508": {
 		id: "mistralai/codestral-2508",
@@ -1700,12 +1718,12 @@ export const OPENROUTER_MODELS = {
 		input: ["text"],
 		cost: {
 			input: 0.02,
-			output: 0.03,
+			output: 0.04,
 			cacheRead: 0,
 			cacheWrite: 0,
 		},
 		contextWindow: 131072,
-		maxTokens: 4096,
+		maxTokens: 16384,
 	} satisfies Model<"openai-completions">,
 	"mistralai/mistral-saba": {
 		id: "mistralai/mistral-saba",
@@ -1753,13 +1771,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: false,
 		input: ["text", "image"],
 		cost: {
-			input: 0.075,
-			output: 0.2,
-			cacheRead: 0,
+			input: 0.1,
+			output: 0.3,
+			cacheRead: 0.01,
 			cacheWrite: 0,
 		},
-		contextWindow: 128000,
-		maxTokens: 16384,
+		contextWindow: 131072,
+		maxTokens: 4096,
 	} satisfies Model<"openai-completions">,
 	"mistralai/mixtral-8x22b-instruct": {
 		id: "mistralai/mixtral-8x22b-instruct",
@@ -1845,11 +1863,11 @@ export const OPENROUTER_MODELS = {
 		cost: {
 			input: 0.6,
 			output: 2.5,
-			cacheRead: 0.15,
+			cacheRead: 0,
 			cacheWrite: 0,
 		},
 		contextWindow: 262144,
-		maxTokens: 100352,
+		maxTokens: 262144,
 	} satisfies Model<"openai-completions">,
 	"moonshotai/kimi-k2.5": {
 		id: "moonshotai/kimi-k2.5",
@@ -1879,13 +1897,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
-			input: 0.66,
-			output: 3.41,
-			cacheRead: 0.15,
+			input: 0.95,
+			output: 4,
+			cacheRead: 0.16,
 			cacheWrite: 0,
 		},
 		contextWindow: 262144,
-		maxTokens: 262144,
+		maxTokens: 4096,
 	} satisfies Model<"openai-completions">,
 	"moonshotai/kimi-k2.7-code": {
 		id: "moonshotai/kimi-k2.7-code",
@@ -1897,9 +1915,9 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
-			input: 0.719,
-			output: 3.49,
-			cacheRead: 0.149,
+			input: 0.75,
+			output: 3.5,
+			cacheRead: 0.16,
 			cacheWrite: 0,
 		},
 		contextWindow: 262144,
@@ -2023,12 +2041,12 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 0.08,
-			output: 0.45,
-			cacheRead: 0,
+			input: 0.21,
+			output: 0.455,
+			cacheRead: 0.06,
 			cacheWrite: 0,
 		},
-		contextWindow: 262144,
+		contextWindow: 1000000,
 		maxTokens: 4096,
 	} satisfies Model<"openai-completions">,
 	"nvidia/nemotron-3-super-120b-a12b:free": {
@@ -2059,13 +2077,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 0.5,
-			output: 2.2,
-			cacheRead: 0.1,
+			input: 0.6,
+			output: 3.6,
+			cacheRead: 0.2,
 			cacheWrite: 0,
 		},
-		contextWindow: 262144,
-		maxTokens: 16384,
+		contextWindow: 512288,
+		maxTokens: 4096,
 	} satisfies Model<"openai-completions">,
 	"nvidia/nemotron-3-ultra-550b-a55b:free": {
 		id: "nvidia/nemotron-3-ultra-550b-a55b:free",
@@ -2245,7 +2263,7 @@ export const OPENROUTER_MODELS = {
 			cacheWrite: 0,
 		},
 		contextWindow: 1047576,
-		maxTokens: 4096,
+		maxTokens: 32768,
 	} satisfies Model<"openai-completions">,
 	"openai/gpt-4.1-mini": {
 		id: "openai/gpt-4.1-mini",
@@ -2295,7 +2313,7 @@ export const OPENROUTER_MODELS = {
 		cost: {
 			input: 2.5,
 			output: 10,
-			cacheRead: 0,
+			cacheRead: 1.25,
 			cacheWrite: 0,
 		},
 		contextWindow: 128000,
@@ -2511,11 +2529,11 @@ export const OPENROUTER_MODELS = {
 		cost: {
 			input: 1.25,
 			output: 10,
-			cacheRead: 0.13,
+			cacheRead: 0.125,
 			cacheWrite: 0,
 		},
 		contextWindow: 128000,
-		maxTokens: 32000,
+		maxTokens: 16384,
 	} satisfies Model<"openai-completions">,
 	"openai/gpt-5.1-codex": {
 		id: "openai/gpt-5.1-codex",
@@ -2529,7 +2547,7 @@ export const OPENROUTER_MODELS = {
 		cost: {
 			input: 1.25,
 			output: 10,
-			cacheRead: 0.13,
+			cacheRead: 0.125,
 			cacheWrite: 0,
 		},
 		contextWindow: 400000,
@@ -2977,8 +2995,8 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 0.03,
-			output: 0.15,
+			input: 0.037,
+			output: 0.17,
 			cacheRead: 0,
 			cacheWrite: 0,
 		},
@@ -2995,13 +3013,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 0.029,
-			output: 0.14,
-			cacheRead: 0,
+			input: 0.03,
+			output: 0.13,
+			cacheRead: 0.03,
 			cacheWrite: 0,
 		},
 		contextWindow: 131072,
-		maxTokens: 4096,
+		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
 	"openai/gpt-oss-20b:free": {
 		id: "openai/gpt-oss-20b:free",
@@ -3517,13 +3535,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: false,
 		input: ["text"],
 		cost: {
-			input: 0.04815,
-			output: 0.19305,
+			input: 0.1,
+			output: 0.3,
 			cacheRead: 0,
 			cacheWrite: 0,
 		},
-		contextWindow: 128000,
-		maxTokens: 32000,
+		contextWindow: 262144,
+		maxTokens: 4096,
 	} satisfies Model<"openai-completions">,
 	"qwen/qwen3-30b-a3b-thinking-2507": {
 		id: "qwen/qwen3-30b-a3b-thinking-2507",
@@ -3589,9 +3607,9 @@ export const OPENROUTER_MODELS = {
 		reasoning: false,
 		input: ["text"],
 		cost: {
-			input: 0.22,
-			output: 1.8,
-			cacheRead: 0,
+			input: 0.3,
+			output: 1,
+			cacheRead: 0.1,
 			cacheWrite: 0,
 		},
 		contextWindow: 262144,
@@ -3733,13 +3751,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: false,
 		input: ["text"],
 		cost: {
-			input: 0.09,
+			input: 0.1,
 			output: 1.1,
-			cacheRead: 0,
+			cacheRead: 0.07,
 			cacheWrite: 0,
 		},
 		contextWindow: 262144,
-		maxTokens: 16384,
+		maxTokens: 262144,
 	} satisfies Model<"openai-completions">,
 	"qwen/qwen3-next-80b-a3b-instruct:free": {
 		id: "qwen/qwen3-next-80b-a3b-instruct:free",
@@ -3787,13 +3805,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: false,
 		input: ["text", "image"],
 		cost: {
-			input: 0.2,
-			output: 0.88,
-			cacheRead: 0.11,
+			input: 0.21,
+			output: 1.9,
+			cacheRead: 0.1,
 			cacheWrite: 0,
 		},
-		contextWindow: 262144,
-		maxTokens: 16384,
+		contextWindow: 131072,
+		maxTokens: 32768,
 	} satisfies Model<"openai-completions">,
 	"qwen/qwen3-vl-235b-a22b-thinking": {
 		id: "qwen/qwen3-vl-235b-a22b-thinking",
@@ -3919,7 +3937,7 @@ export const OPENROUTER_MODELS = {
 			cacheWrite: 0,
 		},
 		contextWindow: 262144,
-		maxTokens: 262144,
+		maxTokens: 65536,
 	} satisfies Model<"openai-completions">,
 	"qwen/qwen3.5-27b": {
 		id: "qwen/qwen3.5-27b",
@@ -3951,11 +3969,11 @@ export const OPENROUTER_MODELS = {
 		cost: {
 			input: 0.14,
 			output: 1,
-			cacheRead: 0.05,
+			cacheRead: 0,
 			cacheWrite: 0,
 		},
 		contextWindow: 262144,
-		maxTokens: 81920,
+		maxTokens: 262144,
 	} satisfies Model<"openai-completions">,
 	"qwen/qwen3.5-397b-a17b": {
 		id: "qwen/qwen3.5-397b-a17b",
@@ -3967,13 +3985,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
-			input: 0.385,
-			output: 2.45,
-			cacheRead: 0.111,
+			input: 0.45,
+			output: 3,
+			cacheRead: 0.225,
 			cacheWrite: 0,
 		},
-		contextWindow: 131072,
-		maxTokens: 4096,
+		contextWindow: 262144,
+		maxTokens: 65536,
 	} satisfies Model<"openai-completions">,
 	"qwen/qwen3.5-9b": {
 		id: "qwen/qwen3.5-9b",
@@ -4057,13 +4075,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
-			input: 0.289,
-			output: 2.4,
+			input: 0.45,
+			output: 2.7,
 			cacheRead: 0,
 			cacheWrite: 0,
 		},
-		contextWindow: 131072,
-		maxTokens: 131072,
+		contextWindow: 262144,
+		maxTokens: 65536,
 	} satisfies Model<"openai-completions">,
 	"qwen/qwen3.6-35b-a3b": {
 		id: "qwen/qwen3.6-35b-a3b",
@@ -4147,10 +4165,10 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 1.25,
-			output: 3.75,
-			cacheRead: 0.25,
-			cacheWrite: 1.5625,
+			input: 1.475,
+			output: 4.425,
+			cacheRead: 0.295,
+			cacheWrite: 1.84375,
 		},
 		contextWindow: 1000000,
 		maxTokens: 65536,
@@ -4291,13 +4309,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 0.14,
-			output: 0.58,
-			cacheRead: 0.035,
+			input: 0.2,
+			output: 0.8,
+			cacheRead: 0.05,
 			cacheWrite: 0,
 		},
 		contextWindow: 262144,
-		maxTokens: 4096,
+		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
 	"tencent/hy3-preview": {
 		id: "tencent/hy3-preview",
@@ -4453,13 +4471,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
-			input: 0.105,
+			input: 0.14,
 			output: 0.28,
-			cacheRead: 0.028,
+			cacheRead: 0.0028,
 			cacheWrite: 0,
 		},
-		contextWindow: 262144,
-		maxTokens: 4096,
+		contextWindow: 1048576,
+		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
 	"xiaomi/mimo-v2.5-pro": {
 		id: "xiaomi/mimo-v2.5-pro",
@@ -4543,13 +4561,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 0.43,
-			output: 1.75,
-			cacheRead: 0.08,
+			input: 0.5,
+			output: 2,
+			cacheRead: 0.1,
 			cacheWrite: 0,
 		},
-		contextWindow: 198000,
-		maxTokens: 16384,
+		contextWindow: 202752,
+		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
 	"z-ai/glm-4.6v": {
 		id: "z-ai/glm-4.6v",
@@ -4597,13 +4615,13 @@ export const OPENROUTER_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 0.06,
+			input: 0.0605,
 			output: 0.4,
-			cacheRead: 0.01,
+			cacheRead: 0,
 			cacheWrite: 0,
 		},
-		contextWindow: 202752,
-		maxTokens: 16384,
+		contextWindow: 131072,
+		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
 	"z-ai/glm-5": {
 		id: "z-ai/glm-5",
@@ -4620,8 +4638,8 @@ export const OPENROUTER_MODELS = {
 			cacheRead: 0.119,
 			cacheWrite: 0,
 		},
-		contextWindow: 198000,
-		maxTokens: 128000,
+		contextWindow: 202752,
+		maxTokens: 202752,
 	} satisfies Model<"openai-completions">,
 	"z-ai/glm-5-turbo": {
 		id: "z-ai/glm-5-turbo",
@@ -4638,7 +4656,7 @@ export const OPENROUTER_MODELS = {
 			cacheRead: 0.24,
 			cacheWrite: 0,
 		},
-		contextWindow: 262144,
+		contextWindow: 202752,
 		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
 	"z-ai/glm-5.1": {
@@ -4670,13 +4688,13 @@ export const OPENROUTER_MODELS = {
 		thinkingLevelMap: {"xhigh":"xhigh"},
 		input: ["text"],
 		cost: {
-			input: 0.924,
-			output: 2.904,
-			cacheRead: 0.1716,
+			input: 0.9366,
+			output: 2.9436,
+			cacheRead: 0.17394,
 			cacheWrite: 0,
 		},
-		contextWindow: 1024000,
-		maxTokens: 128000,
+		contextWindow: 1048576,
+		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
 	"z-ai/glm-5v-turbo": {
 		id: "z-ai/glm-5v-turbo",

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

@@ -622,6 +622,25 @@ export const VERCEL_AI_GATEWAY_MODELS = {
 		contextWindow: 1000000,
 		maxTokens: 128000,
 	} satisfies Model<"anthropic-messages">,
+	"anthropic/claude-opus-4.7-fast": {
+		id: "anthropic/claude-opus-4.7-fast",
+		name: "Claude Opus 4.7 (Fast)",
+		api: "anthropic-messages",
+		provider: "vercel-ai-gateway",
+		baseUrl: "https://ai-gateway.vercel.sh",
+		compat: {"forceAdaptiveThinking":true,"supportsTemperature":false},
+		reasoning: true,
+		thinkingLevelMap: {"xhigh":"xhigh","max":"max"},
+		input: ["text", "image"],
+		cost: {
+			input: 30,
+			output: 150,
+			cacheRead: 3,
+			cacheWrite: 37.5,
+		},
+		contextWindow: 1000000,
+		maxTokens: 128000,
+	} satisfies Model<"anthropic-messages">,
 	"anthropic/claude-opus-4.8": {
 		id: "anthropic/claude-opus-4.8",
 		name: "Claude Opus 4.8",
@@ -641,6 +660,25 @@ export const VERCEL_AI_GATEWAY_MODELS = {
 		contextWindow: 1000000,
 		maxTokens: 128000,
 	} satisfies Model<"anthropic-messages">,
+	"anthropic/claude-opus-4.8-fast": {
+		id: "anthropic/claude-opus-4.8-fast",
+		name: "Claude Opus 4.8 (Fast)",
+		api: "anthropic-messages",
+		provider: "vercel-ai-gateway",
+		baseUrl: "https://ai-gateway.vercel.sh",
+		compat: {"forceAdaptiveThinking":true,"supportsTemperature":false},
+		reasoning: true,
+		thinkingLevelMap: {"xhigh":"xhigh","max":"max"},
+		input: ["text", "image"],
+		cost: {
+			input: 10,
+			output: 50,
+			cacheRead: 1,
+			cacheWrite: 12.5,
+		},
+		contextWindow: 1000000,
+		maxTokens: 128000,
+	} satisfies Model<"anthropic-messages">,
 	"anthropic/claude-sonnet-4": {
 		id: "anthropic/claude-sonnet-4",
 		name: "Claude Sonnet 4",
@@ -841,8 +879,8 @@ export const VERCEL_AI_GATEWAY_MODELS = {
 		reasoning: true,
 		input: ["text"],
 		cost: {
-			input: 0.21,
-			output: 0.79,
+			input: 0.25,
+			output: 0.95,
 			cacheRead: 0.13,
 			cacheWrite: 0,
 		},
@@ -2717,6 +2755,23 @@ export const VERCEL_AI_GATEWAY_MODELS = {
 		contextWindow: 256000,
 		maxTokens: 256000,
 	} satisfies Model<"anthropic-messages">,
+	"thinkingmachines/inkling": {
+		id: "thinkingmachines/inkling",
+		name: "Inkling",
+		api: "anthropic-messages",
+		provider: "vercel-ai-gateway",
+		baseUrl: "https://ai-gateway.vercel.sh",
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 1,
+			output: 4.05,
+			cacheRead: 0.17,
+			cacheWrite: 0,
+		},
+		contextWindow: 256000,
+		maxTokens: 256000,
+	} satisfies Model<"anthropic-messages">,
 	"xai/grok-4.1-fast-non-reasoning": {
 		id: "xai/grok-4.1-fast-non-reasoning",
 		name: "Grok 4.1 Fast Non-Reasoning",

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

@@ -97,11 +97,12 @@ export const XAI_MODELS = {
 	"grok-4.5": {
 		id: "grok-4.5",
 		name: "Grok 4.5",
-		api: "openai-completions",
+		api: "openai-responses",
 		provider: "xai",
 		baseUrl: "https://api.x.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false},
+		compat: {"supportsLongCacheRetention":false},
 		reasoning: true,
+		thinkingLevelMap: {"off":null,"minimal":null},
 		input: ["text", "image"],
 		cost: {
 			input: 2,
@@ -111,7 +112,7 @@ export const XAI_MODELS = {
 		},
 		contextWindow: 500000,
 		maxTokens: 500000,
-	} satisfies Model<"openai-completions">,
+	} satisfies Model<"openai-responses">,
 	"grok-build-0.1": {
 		id: "grok-build-0.1",
 		name: "Grok Build 0.1",

packages/ai/src/providers/xai.ts

@@ -1,15 +1,23 @@
 import { openAICompletionsApi } from "../api/openai-completions.lazy.ts";
-import { envApiKeyAuth } from "../auth/helpers.ts";
+import { openAIResponsesApi } from "../api/openai-responses.lazy.ts";
+import { envApiKeyAuth, lazyOAuth } from "../auth/helpers.ts";
+import { loadXaiOAuth } from "../auth/oauth/load.ts";
 import { createProvider, type Provider } from "../models.ts";
 import { XAI_MODELS } from "./xai.models.ts";
 
-export function xaiProvider(): Provider<"openai-completions"> {
+export function xaiProvider(): Provider<"openai-completions" | "openai-responses"> {
 	return createProvider({
 		id: "xai",
 		name: "xAI",
 		baseUrl: "https://api.x.ai/v1",
-		auth: { apiKey: envApiKeyAuth("xAI API key", ["XAI_API_KEY"]) },
+		auth: {
+			apiKey: envApiKeyAuth("xAI API key", ["XAI_API_KEY"]),
+			oauth: lazyOAuth({ name: "xAI (Grok/X subscription)", load: loadXaiOAuth }),
+		},
 		models: Object.values(XAI_MODELS),
-		api: openAICompletionsApi(),
+		api: {
+			"openai-completions": openAICompletionsApi(),
+			"openai-responses": openAIResponsesApi(),
+		},
 	});
 }