diff --git a/packages/ai/src/provider-models/openai-compat.ts b/packages/ai/src/provider-models/openai-compat.ts
index 7239e6dd5..c20ecd0fd 100644
--- a/packages/ai/src/provider-models/openai-compat.ts
+++ b/packages/ai/src/provider-models/openai-compat.ts
@@ -1410,13 +1410,15 @@ export function xiaomiModelManagerOptions(
 	config?: XiaomiModelManagerConfig,
 ): ModelManagerOptions<"anthropic-messages"> {
 	const apiKey = config?.apiKey;
-	// Xiaomi splits API keys across two backends: standard `sk-` keys hit
-	// api.xiaomimimo.com; "token plan" `tp-` keys hit the EU token-plan host.
-	// Both expose the same Anthropic-compat layout under /anthropic/v1/*.
 	const defaultBaseUrl = apiKey?.startsWith("tp-")
 		? "https://token-plan-ams.xiaomimimo.com/anthropic"
 		: "https://api.xiaomimimo.com/anthropic";
-	const baseUrl = normalizeAnthropicBaseUrl(config?.baseUrl, defaultBaseUrl);
+	// Token-plan keys always use the TP baseUrl; config?.baseUrl (from catalog)
+	// would incorrectly pin to the standard endpoint (api.xiaomimimo.com).
+	const baseUrl = normalizeAnthropicBaseUrl(
+		apiKey?.startsWith("tp-") ? undefined : config?.baseUrl,
+		defaultBaseUrl,
+	);
 	// Xiaomi hosts chat completions under /anthropic/* but exposes model
 	// discovery at the OpenAI-style /v1/models endpoint on the root host.
 	const discoveryRoot = baseUrl.endsWith("/anthropic") ? baseUrl.slice(0, -"/anthropic".length) : baseUrl;
@@ -1439,6 +1441,7 @@ export function xiaomiModelManagerOptions(
 							...model,
 							name: toModelName(entry.display_name, model.name),
 							baseUrl,
+							compat: { ...model.compat, usesApiKeyAuth: true },
 						};
 					},
 				}),
diff --git a/packages/ai/src/providers/anthropic.ts b/packages/ai/src/providers/anthropic.ts
index 9cfd7d310..ea5ecf7dd 100644
--- a/packages/ai/src/providers/anthropic.ts
+++ b/packages/ai/src/providers/anthropic.ts
@@ -80,6 +80,8 @@ export type AnthropicHeaderOptions = {
 	stream?: boolean;
 	modelHeaders?: Record<string, string>;
 	isCloudflareAiGateway?: boolean;
+	/** Use X-Api-Key header instead of Authorization: Bearer. */
+	usesApiKeyAuth?: boolean;
 };
 
 export function normalizeAnthropicBaseUrl(baseUrl?: string): string | undefined {
@@ -182,7 +184,7 @@ export function buildAnthropicHeaders(options: AnthropicHeaderOptions): Record<s
 			"Anthropic-Beta": betaHeader,
 			"User-Agent": userAgent,
 		};
-	} else if (!isAnthropicApiBaseUrl(options.baseUrl)) {
+	} else if (!options.usesApiKeyAuth && !isAnthropicApiBaseUrl(options.baseUrl)) {
 		return {
 			...modelHeaders,
 			Accept: acceptHeader,
@@ -1576,6 +1578,7 @@ export function buildAnthropicClientOptions(args: AnthropicClientOptionsArgs): A
 		stream,
 		modelHeaders: mergeHeaders(model.headers, foundryCustomHeaders, headers, dynamicHeaders),
 		isCloudflareAiGateway: model.provider === "cloudflare-ai-gateway",
+		usesApiKeyAuth: compat.usesApiKeyAuth,
 	});
 
 	if (model.provider === "cloudflare-ai-gateway") {
diff --git a/packages/ai/src/types.ts b/packages/ai/src/types.ts
index f1c8e84ad..ea8dbb30f 100644
--- a/packages/ai/src/types.ts
+++ b/packages/ai/src/types.ts
@@ -791,6 +791,8 @@ export interface AnthropicCompat {
 	supportsEagerToolInputStreaming?: boolean;
 	/** Whether long prompt-cache retention (`ttl: "1h"`) is supported. Default: true for canonical Anthropic API. */
 	supportsLongCacheRetention?: boolean;
+	/** Use X-Api-Key header instead of Authorization: Bearer for non-Anthropic endpoints. */
+	usesApiKeyAuth?: boolean;
 }
 
 /**