feat(ai): complete models runtime migration

2026-06-23 15:29:17 +02:00
parent 470a4736a3
commit 129eb460cd
47 changed files with 1502 additions and 576 deletions
@@ -1,98 +0,0 @@
-import type {
-	Api,
-	AssistantMessageEventStream,
-	Context,
-	Model,
-	SimpleStreamOptions,
-	StreamFunction,
-	StreamOptions,
-} from "./types.ts";
-
-export type ApiStreamFunction = (
-	model: Model<Api>,
-	context: Context,
-	options?: StreamOptions,
-) => AssistantMessageEventStream;
-
-export type ApiStreamSimpleFunction = (
-	model: Model<Api>,
-	context: Context,
-	options?: SimpleStreamOptions,
-) => AssistantMessageEventStream;
-
-export interface ApiProvider<TApi extends Api = Api, TOptions extends StreamOptions = StreamOptions> {
-	api: TApi;
-	stream: StreamFunction<TApi, TOptions>;
-	streamSimple: StreamFunction<TApi, SimpleStreamOptions>;
-}
-
-interface ApiProviderInternal {
-	api: Api;
-	stream: ApiStreamFunction;
-	streamSimple: ApiStreamSimpleFunction;
-}
-
-type RegisteredApiProvider = {
-	provider: ApiProviderInternal;
-	sourceId?: string;
-};
-
-const apiProviderRegistry = new Map<string, RegisteredApiProvider>();
-
-function wrapStream<TApi extends Api, TOptions extends StreamOptions>(
-	api: TApi,
-	stream: StreamFunction<TApi, TOptions>,
-): ApiStreamFunction {
-	return (model, context, options) => {
-		if (model.api !== api) {
-			throw new Error(`Mismatched api: ${model.api} expected ${api}`);
-		}
-		return stream(model as Model<TApi>, context, options as TOptions);
-	};
-}
-
-function wrapStreamSimple<TApi extends Api>(
-	api: TApi,
-	streamSimple: StreamFunction<TApi, SimpleStreamOptions>,
-): ApiStreamSimpleFunction {
-	return (model, context, options) => {
-		if (model.api !== api) {
-			throw new Error(`Mismatched api: ${model.api} expected ${api}`);
-		}
-		return streamSimple(model as Model<TApi>, context, options);
-	};
-}
-
-export function registerApiProvider<TApi extends Api, TOptions extends StreamOptions>(
-	provider: ApiProvider<TApi, TOptions>,
-	sourceId?: string,
-): void {
-	apiProviderRegistry.set(provider.api, {
-		provider: {
-			api: provider.api,
-			stream: wrapStream(provider.api, provider.stream),
-			streamSimple: wrapStreamSimple(provider.api, provider.streamSimple),
-		},
-		sourceId,
-	});
-}
-
-export function getApiProvider(api: Api): ApiProviderInternal | undefined {
-	return apiProviderRegistry.get(api)?.provider;
-}
-
-export function getApiProviders(): ApiProviderInternal[] {
-	return Array.from(apiProviderRegistry.values(), (entry) => entry.provider);
-}
-
-export function unregisterApiProviders(sourceId: string): void {
-	for (const [api, entry] of apiProviderRegistry.entries()) {
-		if (entry.sourceId === sourceId) {
-			apiProviderRegistry.delete(api);
-		}
-	}
-}
-
-export function clearApiProviders(): void {
-	apiProviderRegistry.clear();
-}
@@ -18,6 +18,7 @@ import type {
 	Message,
 	Model,
 	ProviderEnv,
+	ProviderHeaders,
 	SimpleStreamOptions,
 	StopReason,
 	StreamFunction,
@@ -34,7 +35,6 @@ import { parseJsonWithRepair, parseStreamingJson } from "../utils/json-parse.ts"
 import { getProviderEnvValue } from "../utils/provider-env.ts";
 import { sanitizeSurrogates } from "../utils/sanitize-unicode.ts";

-import { resolveCloudflareBaseUrl } from "./cloudflare.ts";
 import { buildCopilotDynamicHeaders, hasCopilotVisionInput } from "./github-copilot-headers.ts";
 import { adjustMaxTokensForThinking, buildBaseOptions } from "./simple-options.ts";
 import { transformMessages } from "./transform-messages.ts";
@@ -170,16 +170,11 @@ const INTERLEAVED_THINKING_BETA = "interleaved-thinking-2025-05-14";
 function getAnthropicCompat(
 	model: Model<"anthropic-messages">,
 ): Required<Omit<AnthropicMessagesCompat, "forceAdaptiveThinking">> {
-	// Auto-detect session affinity and cache control support from provider
-	const isFireworks = model.provider === "fireworks";
-	const isCloudflareAiGatewayAnthropic =
-		model.provider === "cloudflare-ai-gateway" && model.baseUrl.includes("anthropic");
 	return {
-		supportsEagerToolInputStreaming: model.compat?.supportsEagerToolInputStreaming ?? !isFireworks,
-		supportsLongCacheRetention: model.compat?.supportsLongCacheRetention ?? !isFireworks,
-		sendSessionAffinityHeaders:
-			model.compat?.sendSessionAffinityHeaders ?? !!(isFireworks || isCloudflareAiGatewayAnthropic),
-		supportsCacheControlOnTools: model.compat?.supportsCacheControlOnTools ?? !isFireworks,
+		supportsEagerToolInputStreaming: model.compat?.supportsEagerToolInputStreaming ?? true,
+		supportsLongCacheRetention: model.compat?.supportsLongCacheRetention ?? true,
+		sendSessionAffinityHeaders: model.compat?.sendSessionAffinityHeaders ?? false,
+		supportsCacheControlOnTools: model.compat?.supportsCacheControlOnTools ?? true,
 		supportsTemperature: model.compat?.supportsTemperature ?? true,
 		allowEmptySignature: model.compat?.allowEmptySignature ?? false,
 	};
@@ -247,8 +242,8 @@ export interface AnthropicOptions extends StreamOptions {
 	client?: Anthropic;
 }

-function mergeHeaders(...headerSources: (Record<string, string | null> | undefined)[]): Record<string, string | null> {
-	const merged: Record<string, string | null> = {};
+function mergeHeaders(...headerSources: (ProviderHeaders | undefined)[]): ProviderHeaders {
+	const merged: ProviderHeaders = {};
 	for (const headers of headerSources) {
 		if (headers) {
 			Object.assign(merged, headers);
@@ -257,6 +252,27 @@ function mergeHeaders(...headerSources: (Record<string, string | null> | undefin
 	return merged;
 }

+function hasHeader(headers: ProviderHeaders | undefined, name: string): boolean {
+	if (!headers) return false;
+	const expected = name.toLowerCase();
+	for (const [key, value] of Object.entries(headers)) {
+		if (key.toLowerCase() === expected && value !== null && value.trim().length > 0) return true;
+	}
+	return false;
+}
+
+function assertRequestAuth(provider: string, apiKey: string | undefined, headers: ProviderHeaders | undefined): void {
+	if (apiKey) return;
+	if (
+		hasHeader(headers, "authorization") ||
+		hasHeader(headers, "x-api-key") ||
+		hasHeader(headers, "cf-aig-authorization")
+	) {
+		return;
+	}
+	throw new Error(`No API key for provider: ${provider}`);
+}
+
 interface ServerSentEvent {
 	event: string | null;
 	data: string;
@@ -484,9 +500,7 @@ export const stream: StreamFunction<"anthropic-messages", AnthropicOptions> = (
 				isOAuth = false;
 			} else {
 				const apiKey = options?.apiKey;
-				if (!apiKey) {
-					throw new Error(`No API key for provider: ${model.provider}`);
-				}
+				assertRequestAuth(model.provider, apiKey, options?.headers);

 				let copilotDynamicHeaders: Record<string, string> | undefined;
 				if (model.provider === "github-copilot") {
@@ -508,7 +522,6 @@ export const stream: StreamFunction<"anthropic-messages", AnthropicOptions> = (
 					options?.headers,
 					copilotDynamicHeaders,
 					cacheSessionId,
-					options?.env,
 				);
 				client = created.client;
 				isOAuth = created.isOAuthToken;
@@ -748,12 +761,9 @@ export const streamSimple: StreamFunction<"anthropic-messages", SimpleStreamOpti
 	context: Context,
 	options?: SimpleStreamOptions,
 ): AssistantMessageEventStream => {
-	const apiKey = options?.apiKey;
-	if (!apiKey) {
-		throw new Error(`No API key for provider: ${model.provider}`);
-	}
+	assertRequestAuth(model.provider, options?.apiKey, options?.headers);

-	const base = buildBaseOptions(model, options, apiKey);
+	const base = buildBaseOptions(model, options, options?.apiKey);
 	if (!options?.reasoning) {
 		return stream(model, context, { ...base, thinkingEnabled: false } satisfies AnthropicOptions);
 	}
@@ -792,13 +802,12 @@ function isOAuthToken(apiKey: string): boolean {

 function createClient(
 	model: Model<"anthropic-messages">,
-	apiKey: string,
+	apiKey: string | undefined,
 	interleavedThinking: boolean,
 	useFineGrainedToolStreamingBeta: boolean,
-	optionsHeaders?: Record<string, string>,
+	optionsHeaders?: ProviderHeaders,
 	dynamicHeaders?: Record<string, string>,
 	sessionId?: string,
-	env?: ProviderEnv,
 ): { client: Anthropic; isOAuthToken: boolean } {
 	// Adaptive thinking models have interleaved thinking built in, so skip the beta header.
 	const needsInterleavedBeta = interleavedThinking && model.compat?.forceAdaptiveThinking !== true;
@@ -810,34 +819,11 @@ function createClient(
 		betaFeatures.push(INTERLEAVED_THINKING_BETA);
 	}

-	if (model.provider === "cloudflare-ai-gateway") {
-		const client = new Anthropic({
-			apiKey: null,
-			authToken: null,
-			baseURL: resolveCloudflareBaseUrl(model, env),
-			dangerouslyAllowBrowser: true,
-			defaultHeaders: mergeHeaders(
-				{
-					accept: "application/json",
-					"anthropic-dangerous-direct-browser-access": "true",
-					"cf-aig-authorization": `Bearer ${apiKey}`,
-					"x-api-key": null,
-					Authorization: null,
-					...(betaFeatures.length > 0 ? { "anthropic-beta": betaFeatures.join(",") } : {}),
-				},
-				model.headers,
-				optionsHeaders,
-			),
-		});
-
-		return { client, isOAuthToken: false };
-	}
-
 	// Copilot: Bearer auth, selective betas.
 	if (model.provider === "github-copilot") {
 		const client = new Anthropic({
 			apiKey: null,
-			authToken: apiKey,
+			authToken: apiKey ?? null,
 			baseURL: model.baseUrl,
 			dangerouslyAllowBrowser: true,
 			defaultHeaders: mergeHeaders(
@@ -856,7 +842,7 @@ function createClient(
 	}

 	// OAuth: Bearer auth, Claude Code identity headers
-	if (isOAuthToken(apiKey)) {
+	if (apiKey && isOAuthToken(apiKey)) {
 		const client = new Anthropic({
 			apiKey: null,
 			authToken: apiKey,
@@ -878,24 +864,25 @@ function createClient(
 		return { client, isOAuthToken: true };
 	}

-	// API key auth
-	const sessionAffinityHeaders: Record<string, string | null> =
+	// API key or header-owned auth.
+	const sessionAffinityHeaders: ProviderHeaders =
 		sessionId && getAnthropicCompat(model).sendSessionAffinityHeaders ? { "x-session-affinity": sessionId } : {};
+	const defaultHeaders = mergeHeaders(
+		{
+			accept: "application/json",
+			"anthropic-dangerous-direct-browser-access": "true",
+			...(betaFeatures.length > 0 ? { "anthropic-beta": betaFeatures.join(",") } : {}),
+		},
+		sessionAffinityHeaders,
+		model.headers,
+		optionsHeaders,
+	);
 	const client = new Anthropic({
-		apiKey,
+		apiKey: apiKey ?? null,
 		authToken: null,
 		baseURL: model.baseUrl,
 		dangerouslyAllowBrowser: true,
-		defaultHeaders: mergeHeaders(
-			{
-				accept: "application/json",
-				"anthropic-dangerous-direct-browser-access": "true",
-				...(betaFeatures.length > 0 ? { "anthropic-beta": betaFeatures.join(",") } : {}),
-			},
-			sessionAffinityHeaders,
-			model.headers,
-			optionsHeaders,
-		),
+		defaultHeaders,
 	});

 	return { client, isOAuthToken: false };
@@ -48,6 +48,7 @@ import type {
 	ToolResultMessage,
 } from "../types.ts";
 import { AssistantMessageEventStream } from "../utils/event-stream.ts";
+import { providerHeadersToRecord } from "../utils/headers.ts";
 import { parseStreamingJson } from "../utils/json-parse.ts";
 import { resolveHttpProxyUrlForTarget } from "../utils/node-http-proxy.ts";
 import { getProviderEnvValue } from "../utils/provider-env.ts";
@@ -204,8 +205,9 @@ export const stream: StreamFunction<"bedrock-converse-stream", BedrockOptions> =

 		try {
 			const client = new BedrockRuntimeClient(config);
-			if (options.headers && Object.keys(options.headers).length > 0) {
-				addCustomHeadersMiddleware(client, options.headers);
+			const customHeaders = providerHeadersToRecord(options.headers);
+			if (customHeaders) {
+				addCustomHeadersMiddleware(client, customHeaders);
 			}
 			const cacheRetention = resolveCacheRetention(options.cacheRetention, options.env);
 			const inferenceMaxTokens = options.maxTokens ?? (isAnthropicClaudeModel(model) ? model.maxTokens : undefined);
@@ -1,6 +1,3 @@
-import type { Api, Model, ProviderEnv } from "../types.ts";
-import { getProviderEnvValue } from "../utils/provider-env.ts";
-
 /** Workers AI direct endpoint. */
 export const CLOUDFLARE_WORKERS_AI_BASE_URL =
 	"https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1";
@@ -16,21 +13,3 @@ export const CLOUDFLARE_AI_GATEWAY_OPENAI_BASE_URL =
 /** AI Gateway → Anthropic passthrough. */
 export const CLOUDFLARE_AI_GATEWAY_ANTHROPIC_BASE_URL =
 	"https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic";
-
-export function isCloudflareProvider(provider: string): boolean {
-	return provider === "cloudflare-workers-ai" || provider === "cloudflare-ai-gateway";
-}
-
-/** Substitute `{VAR}` placeholders in a Cloudflare baseUrl from provider env or process.env. */
-export function resolveCloudflareBaseUrl(model: Model<Api>, env?: ProviderEnv): string {
-	const url = model.baseUrl;
-	if (!url.includes("{")) return url;
-	const baseUrl = url.replace(/\{([A-Z_][A-Z0-9_]*)\}/g, (_match, name: string) => {
-		const value = getProviderEnvValue(name, env);
-		if (!value) {
-			throw new Error(`${name} is required for provider ${model.provider} but is not set.`);
-		}
-		return value;
-	});
-	return baseUrl;
-}
@@ -10,6 +10,7 @@ import type {
 	AssistantMessage,
 	Context,
 	Model,
+	ProviderHeaders,
 	SimpleStreamOptions,
 	StreamFunction,
 	StreamOptions,
@@ -20,6 +21,7 @@ import type {
 	ToolCall,
 } from "../types.ts";
 import { AssistantMessageEventStream } from "../utils/event-stream.ts";
+import { providerHeadersToRecord } from "../utils/headers.ts";
 import { sanitizeSurrogates } from "../utils/sanitize-unicode.ts";
 import type { GoogleThinkingLevel } from "./google-shared.ts";
 import {
@@ -318,15 +320,16 @@ export const streamSimple: StreamFunction<"google-generative-ai", SimpleStreamOp
 function createClient(
 	model: Model<"google-generative-ai">,
 	apiKey?: string,
-	optionsHeaders?: Record<string, string>,
+	optionsHeaders?: ProviderHeaders,
 ): GoogleGenAI {
 	const httpOptions: { baseUrl?: string; apiVersion?: string; headers?: Record<string, string> } = {};
 	if (model.baseUrl) {
 		httpOptions.baseUrl = model.baseUrl;
 		httpOptions.apiVersion = ""; // baseUrl already includes version path, don't append
 	}
-	if (model.headers || optionsHeaders) {
-		httpOptions.headers = { ...model.headers, ...optionsHeaders };
+	const headers = providerHeadersToRecord({ ...model.headers, ...optionsHeaders });
+	if (headers) {
+		httpOptions.headers = headers;
 	}

 	return new GoogleGenAI({
@@ -15,6 +15,7 @@ import type {
 	Model,
 	ThinkingLevel as PiThinkingLevel,
 	ProviderEnv,
+	ProviderHeaders,
 	SimpleStreamOptions,
 	StreamFunction,
 	StreamOptions,
@@ -24,6 +25,7 @@ import type {
 	ToolCall,
 } from "../types.ts";
 import { AssistantMessageEventStream } from "../utils/event-stream.ts";
+import { providerHeadersToRecord } from "../utils/headers.ts";
 import { getProviderEnvValue } from "../utils/provider-env.ts";
 import { sanitizeSurrogates } from "../utils/sanitize-unicode.ts";
 import type { GoogleThinkingLevel } from "./google-shared.ts";
@@ -334,7 +336,7 @@ function createClient(
 	model: Model<"google-vertex">,
 	project: string,
 	location: string,
-	optionsHeaders?: Record<string, string>,
+	optionsHeaders?: ProviderHeaders,
 	env?: ProviderEnv,
 ): GoogleGenAI {
 	const googleAuthOptions = buildGoogleAuthOptions(env);
@@ -351,7 +353,7 @@ function createClient(
 function createClientWithApiKey(
 	model: Model<"google-vertex">,
 	apiKey: string,
-	optionsHeaders?: Record<string, string>,
+	optionsHeaders?: ProviderHeaders,
 ): GoogleGenAI {
 	return new GoogleGenAI({
 		vertexai: true,
@@ -361,10 +363,7 @@ function createClientWithApiKey(
 	});
 }

-function buildHttpOptions(
-	model: Model<"google-vertex">,
-	optionsHeaders?: Record<string, string>,
-): HttpOptions | undefined {
+function buildHttpOptions(model: Model<"google-vertex">, optionsHeaders?: ProviderHeaders): HttpOptions | undefined {
 	const httpOptions: HttpOptions = {};
 	const baseUrl = resolveCustomBaseUrl(model.baseUrl);
 	if (baseUrl) {
@@ -375,8 +374,9 @@ function buildHttpOptions(
 		}
 	}

-	if (model.headers || optionsHeaders) {
-		httpOptions.headers = { ...model.headers, ...optionsHeaders };
+	const headers = providerHeadersToRecord({ ...model.headers, ...optionsHeaders });
+	if (headers) {
+		httpOptions.headers = headers;
 	}

 	return Object.keys(httpOptions).length > 0 ? httpOptions : undefined;
@@ -28,6 +28,7 @@ import type {
 	Context,
 	Model,
 	ProviderEnv,
+	ProviderHeaders,
 	SimpleStreamOptions,
 	StreamFunction,
 	StreamOptions,
@@ -1467,13 +1468,17 @@ function createCodexRequestId(): string {

 function buildBaseCodexHeaders(
 	initHeaders: Record<string, string> | undefined,
-	additionalHeaders: Record<string, string> | undefined,
+	additionalHeaders: ProviderHeaders | undefined,
 	accountId: string,
 	token: string,
 ): Headers {
 	const headers = new Headers(initHeaders);
 	for (const [key, value] of Object.entries(additionalHeaders || {})) {
-		headers.set(key, value);
+		if (value === null) {
+			headers.delete(key);
+		} else {
+			headers.set(key, value);
+		}
 	}
 	headers.set("Authorization", `Bearer ${token}`);
 	headers.set("chatgpt-account-id", accountId);
@@ -1485,7 +1490,7 @@ function buildBaseCodexHeaders(

 function buildSSEHeaders(
 	initHeaders: Record<string, string> | undefined,
-	additionalHeaders: Record<string, string> | undefined,
+	additionalHeaders: ProviderHeaders | undefined,
 	accountId: string,
 	token: string,
 	sessionId?: string,
@@ -1505,7 +1510,7 @@ function buildSSEHeaders(

 function buildWebSocketHeaders(
 	initHeaders: Record<string, string> | undefined,
-	additionalHeaders: Record<string, string> | undefined,
+	additionalHeaders: ProviderHeaders | undefined,
 	accountId: string,
 	token: string,
 	requestId: string,
@@ -21,6 +21,7 @@ import type {
 	Model,
 	OpenAICompletionsCompat,
 	ProviderEnv,
+	ProviderHeaders,
 	SimpleStreamOptions,
 	StopReason,
 	StreamFunction,
@@ -36,7 +37,6 @@ import { headersToRecord } from "../utils/headers.ts";
 import { parseStreamingJson } from "../utils/json-parse.ts";
 import { getProviderEnvValue } from "../utils/provider-env.ts";
 import { sanitizeSurrogates } from "../utils/sanitize-unicode.ts";
-import { isCloudflareProvider, resolveCloudflareBaseUrl } from "./cloudflare.ts";
 import { buildCopilotDynamicHeaders, hasCopilotVisionInput } from "./github-copilot-headers.ts";
 import { clampOpenAIPromptCacheKey } from "./openai-prompt-cache.ts";
 import { buildBaseOptions } from "./simple-options.ts";
@@ -47,6 +47,21 @@ import { transformMessages } from "./transform-messages.ts";
 * This is needed because Anthropic (via proxy) requires the tools param
 * to be present when messages include tool_calls or tool role messages.
 */
+function hasHeader(headers: ProviderHeaders | undefined, name: string): boolean {
+	if (!headers) return false;
+	const expected = name.toLowerCase();
+	for (const [key, value] of Object.entries(headers)) {
+		if (key.toLowerCase() === expected && value !== null && value.trim().length > 0) return true;
+	}
+	return false;
+}
+
+function getClientApiKey(provider: string, apiKey: string | undefined, headers: ProviderHeaders | undefined): string {
+	if (apiKey) return apiKey;
+	if (hasHeader(headers, "authorization") || hasHeader(headers, "cf-aig-authorization")) return "unused";
+	throw new Error(`No API key for provider: ${provider}`);
+}
+
 function hasToolHistory(messages: Message[]): boolean {
 	for (const msg of messages) {
 		if (msg.role === "toolResult") {
@@ -160,14 +175,11 @@ export const stream: StreamFunction<"openai-completions", OpenAICompletionsOptio
 		};

 		try {
-			const apiKey = options?.apiKey;
-			if (!apiKey) {
-				throw new Error(`No API key for provider: ${model.provider}`);
-			}
+			const apiKey = getClientApiKey(model.provider, options?.apiKey, options?.headers);
 			const compat = getCompat(model);
 			const cacheRetention = resolveCacheRetention(options?.cacheRetention, options?.env);
 			const cacheSessionId = cacheRetention === "none" ? undefined : options?.sessionId;
-			const client = createClient(model, context, apiKey, options?.headers, cacheSessionId, compat, options?.env);
+			const client = createClient(model, context, apiKey, options?.headers, cacheSessionId, compat);
 			let params = buildParams(model, context, options, compat, cacheRetention);
 			const nextParams = await options?.onPayload?.(params, model);
 			if (nextParams !== undefined) {
@@ -468,12 +480,9 @@ export const streamSimple: StreamFunction<"openai-completions", SimpleStreamOpti
 	context: Context,
 	options?: SimpleStreamOptions,
 ): AssistantMessageEventStream => {
-	const apiKey = options?.apiKey;
-	if (!apiKey) {
-		throw new Error(`No API key for provider: ${model.provider}`);
-	}
+	getClientApiKey(model.provider, options?.apiKey, options?.headers);

-	const base = buildBaseOptions(model, options, apiKey);
+	const base = buildBaseOptions(model, options, options?.apiKey);
 	const clampedReasoning = options?.reasoning ? clampThinkingLevel(model, options.reasoning) : undefined;
 	const reasoningEffort = clampedReasoning === "off" ? undefined : clampedReasoning;
 	const toolChoice = (options as OpenAICompletionsOptions | undefined)?.toolChoice;
@@ -489,12 +498,11 @@ function createClient(
 	model: Model<"openai-completions">,
 	context: Context,
 	apiKey: string,
-	optionsHeaders?: Record<string, string>,
+	optionsHeaders?: ProviderHeaders,
 	sessionId?: string,
 	compat: ResolvedOpenAICompletionsCompat = getCompat(model),
-	env?: ProviderEnv,
 ) {
-	const headers = { ...model.headers };
+	const headers: ProviderHeaders = { ...model.headers };
 	if (model.provider === "github-copilot") {
 		const hasImages = hasCopilotVisionInput(context.messages);
 		const copilotHeaders = buildCopilotDynamicHeaders({
@@ -515,20 +523,11 @@ function createClient(
 		Object.assign(headers, optionsHeaders);
 	}

-	const defaultHeaders =
-		model.provider === "cloudflare-ai-gateway"
-			? {
-					...headers,
-					Authorization: headers.Authorization ?? null,
-					"cf-aig-authorization": `Bearer ${apiKey}`,
-				}
-			: headers;
-
 	return new OpenAI({
 		apiKey,
-		baseURL: isCloudflareProvider(model.provider) ? resolveCloudflareBaseUrl(model, env) : model.baseUrl,
+		baseURL: model.baseUrl,
 		dangerouslyAllowBrowser: true,
-		defaultHeaders,
+		defaultHeaders: headers,
 	});
 }

@@ -674,7 +673,7 @@ function buildParams(
 	}

 	// Vercel AI Gateway provider routing preferences
-	if (model.baseUrl.includes("ai-gateway.vercel.sh") && model.compat?.vercelGatewayRouting) {
+	if (model.compat?.vercelGatewayRouting) {
 		const routing = model.compat.vercelGatewayRouting;
 		if (routing.only || routing.order) {
 			const gatewayOptions: Record<string, string[]> = {};
@@ -1164,123 +1163,55 @@ function mapStopReason(reason: ChatCompletionChunk.Choice["finish_reason"] | str
 	}
 }

-/**
- * Detect compatibility settings from provider and baseUrl for known providers.
- * Provider takes precedence over URL-based detection since it's explicitly configured.
- * Returns a fully resolved OpenAICompletionsCompat object with all fields set.
- */
-function detectCompat(model: Model<"openai-completions">): ResolvedOpenAICompletionsCompat {
-	const provider = model.provider;
-	const baseUrl = model.baseUrl;
-
-	const isZai =
-		provider === "zai" ||
-		provider === "zai-coding-cn" ||
-		baseUrl.includes("api.z.ai") ||
-		baseUrl.includes("open.bigmodel.cn");
-	const isTogether =
-		provider === "together" || baseUrl.includes("api.together.ai") || baseUrl.includes("api.together.xyz");
-	const isMoonshot = provider === "moonshotai" || provider === "moonshotai-cn" || baseUrl.includes("api.moonshot.");
-	const isOpenRouter = provider === "openrouter" || baseUrl.includes("openrouter.ai");
-	const isCloudflareWorkersAI = provider === "cloudflare-workers-ai" || baseUrl.includes("api.cloudflare.com");
-	const isCloudflareAiGateway = provider === "cloudflare-ai-gateway" || baseUrl.includes("gateway.ai.cloudflare.com");
-	const isNvidia = provider === "nvidia" || baseUrl.includes("integrate.api.nvidia.com");
-	const isAntLing = provider === "ant-ling" || baseUrl.includes("api.ant-ling.com");
-
-	const isNonStandard =
-		isNvidia ||
-		provider === "cerebras" ||
-		baseUrl.includes("cerebras.ai") ||
-		provider === "xai" ||
-		baseUrl.includes("api.x.ai") ||
-		isTogether ||
-		baseUrl.includes("chutes.ai") ||
-		baseUrl.includes("deepseek.com") ||
-		isZai ||
-		isMoonshot ||
-		provider === "opencode" ||
-		baseUrl.includes("opencode.ai") ||
-		isCloudflareWorkersAI ||
-		isCloudflareAiGateway ||
-		isAntLing;
-
-	const useMaxTokens =
-		baseUrl.includes("chutes.ai") || isMoonshot || isCloudflareAiGateway || isTogether || isNvidia || isAntLing;
-
-	const isGrok = provider === "xai" || baseUrl.includes("api.x.ai");
-	const isDeepSeek = provider === "deepseek" || baseUrl.includes("deepseek.com");
-	const isOpenRouterDeveloperRoleModel =
-		isOpenRouter && (model.id.startsWith("anthropic/") || model.id.startsWith("openai/"));
-	const cacheControlFormat = provider === "openrouter" && model.id.startsWith("anthropic/") ? "anthropic" : undefined;
-
-	return {
-		supportsStore: !isNonStandard,
-		supportsDeveloperRole: isOpenRouterDeveloperRoleModel || (!isNonStandard && !isOpenRouter),
-		supportsReasoningEffort:
-			!isGrok && !isZai && !isMoonshot && !isTogether && !isCloudflareAiGateway && !isNvidia && !isAntLing,
-		supportsUsageInStreaming: true,
-		maxTokensField: useMaxTokens ? "max_tokens" : "max_completion_tokens",
-		requiresToolResultName: false,
-		requiresAssistantAfterToolResult: false,
-		requiresThinkingAsText: false,
-		requiresReasoningContentOnAssistantMessages: isDeepSeek,
-		thinkingFormat: isDeepSeek
-			? "deepseek"
-			: isZai
-				? "zai"
-				: isTogether
-					? "together"
-					: isAntLing
-						? "ant-ling"
-						: isOpenRouter
-							? "openrouter"
-							: "openai",
-		openRouterRouting: {},
-		vercelGatewayRouting: {},
-		chatTemplateKwargs: {},
-		zaiToolStream: false,
-		supportsStrictMode: !isMoonshot && !isTogether && !isCloudflareAiGateway && !isNvidia,
-		cacheControlFormat,
-		sendSessionAffinityHeaders: false,
-		supportsLongCacheRetention: !(
-			isTogether ||
-			isCloudflareWorkersAI ||
-			isCloudflareAiGateway ||
-			isNvidia ||
-			isAntLing
-		),
-	};
-}
+const DEFAULT_COMPAT: ResolvedOpenAICompletionsCompat = {
+	supportsStore: true,
+	supportsDeveloperRole: true,
+	supportsReasoningEffort: true,
+	supportsUsageInStreaming: true,
+	maxTokensField: "max_completion_tokens",
+	requiresToolResultName: false,
+	requiresAssistantAfterToolResult: false,
+	requiresThinkingAsText: false,
+	requiresReasoningContentOnAssistantMessages: false,
+	thinkingFormat: "openai",
+	openRouterRouting: {},
+	vercelGatewayRouting: {},
+	chatTemplateKwargs: {},
+	zaiToolStream: false,
+	supportsStrictMode: true,
+	cacheControlFormat: undefined,
+	sendSessionAffinityHeaders: false,
+	supportsLongCacheRetention: true,
+};

 /**
 * Get resolved compatibility settings for a model.
- * Uses explicit model.compat if provided, otherwise auto-detects from provider/URL.
+ * Uses explicit generated/custom model.compat over OpenAI-standard defaults.
 */
 function getCompat(model: Model<"openai-completions">): ResolvedOpenAICompletionsCompat {
-	const detected = detectCompat(model);
-	if (!model.compat) return detected;
+	if (!model.compat) return DEFAULT_COMPAT;

 	return {
-		supportsStore: model.compat.supportsStore ?? detected.supportsStore,
-		supportsDeveloperRole: model.compat.supportsDeveloperRole ?? detected.supportsDeveloperRole,
-		supportsReasoningEffort: model.compat.supportsReasoningEffort ?? detected.supportsReasoningEffort,
-		supportsUsageInStreaming: model.compat.supportsUsageInStreaming ?? detected.supportsUsageInStreaming,
-		maxTokensField: model.compat.maxTokensField ?? detected.maxTokensField,
-		requiresToolResultName: model.compat.requiresToolResultName ?? detected.requiresToolResultName,
+		supportsStore: model.compat.supportsStore ?? DEFAULT_COMPAT.supportsStore,
+		supportsDeveloperRole: model.compat.supportsDeveloperRole ?? DEFAULT_COMPAT.supportsDeveloperRole,
+		supportsReasoningEffort: model.compat.supportsReasoningEffort ?? DEFAULT_COMPAT.supportsReasoningEffort,
+		supportsUsageInStreaming: model.compat.supportsUsageInStreaming ?? DEFAULT_COMPAT.supportsUsageInStreaming,
+		maxTokensField: model.compat.maxTokensField ?? DEFAULT_COMPAT.maxTokensField,
+		requiresToolResultName: model.compat.requiresToolResultName ?? DEFAULT_COMPAT.requiresToolResultName,
 		requiresAssistantAfterToolResult:
-			model.compat.requiresAssistantAfterToolResult ?? detected.requiresAssistantAfterToolResult,
-		requiresThinkingAsText: model.compat.requiresThinkingAsText ?? detected.requiresThinkingAsText,
+			model.compat.requiresAssistantAfterToolResult ?? DEFAULT_COMPAT.requiresAssistantAfterToolResult,
+		requiresThinkingAsText: model.compat.requiresThinkingAsText ?? DEFAULT_COMPAT.requiresThinkingAsText,
 		requiresReasoningContentOnAssistantMessages:
 			model.compat.requiresReasoningContentOnAssistantMessages ??
-			detected.requiresReasoningContentOnAssistantMessages,
-		thinkingFormat: model.compat.thinkingFormat ?? detected.thinkingFormat,
-		openRouterRouting: model.compat.openRouterRouting ?? {},
-		vercelGatewayRouting: model.compat.vercelGatewayRouting ?? detected.vercelGatewayRouting,
-		chatTemplateKwargs: model.compat.chatTemplateKwargs ?? detected.chatTemplateKwargs,
-		zaiToolStream: model.compat.zaiToolStream ?? detected.zaiToolStream,
-		supportsStrictMode: model.compat.supportsStrictMode ?? detected.supportsStrictMode,
-		cacheControlFormat: model.compat.cacheControlFormat ?? detected.cacheControlFormat,
-		sendSessionAffinityHeaders: model.compat.sendSessionAffinityHeaders ?? detected.sendSessionAffinityHeaders,
-		supportsLongCacheRetention: model.compat.supportsLongCacheRetention ?? detected.supportsLongCacheRetention,
+			DEFAULT_COMPAT.requiresReasoningContentOnAssistantMessages,
+		thinkingFormat: model.compat.thinkingFormat ?? DEFAULT_COMPAT.thinkingFormat,
+		openRouterRouting: model.compat.openRouterRouting ?? DEFAULT_COMPAT.openRouterRouting,
+		vercelGatewayRouting: model.compat.vercelGatewayRouting ?? DEFAULT_COMPAT.vercelGatewayRouting,
+		chatTemplateKwargs: model.compat.chatTemplateKwargs ?? DEFAULT_COMPAT.chatTemplateKwargs,
+		zaiToolStream: model.compat.zaiToolStream ?? DEFAULT_COMPAT.zaiToolStream,
+		supportsStrictMode: model.compat.supportsStrictMode ?? DEFAULT_COMPAT.supportsStrictMode,
+		cacheControlFormat: model.compat.cacheControlFormat ?? DEFAULT_COMPAT.cacheControlFormat,
+		sendSessionAffinityHeaders: model.compat.sendSessionAffinityHeaders ?? DEFAULT_COMPAT.sendSessionAffinityHeaders,
+		supportsLongCacheRetention: model.compat.supportsLongCacheRetention ?? DEFAULT_COMPAT.supportsLongCacheRetention,
 	};
 }
@@ -9,6 +9,7 @@ import type {
 	Model,
 	OpenAIResponsesCompat,
 	ProviderEnv,
+	ProviderHeaders,
 	SimpleStreamOptions,
 	StreamFunction,
 	StreamOptions,
@@ -17,7 +18,6 @@ import type {
 import { AssistantMessageEventStream } from "../utils/event-stream.ts";
 import { headersToRecord } from "../utils/headers.ts";
 import { getProviderEnvValue } from "../utils/provider-env.ts";
-import { isCloudflareProvider, resolveCloudflareBaseUrl } from "./cloudflare.ts";
 import { buildCopilotDynamicHeaders, hasCopilotVisionInput } from "./github-copilot-headers.ts";
 import { clampOpenAIPromptCacheKey } from "./openai-prompt-cache.ts";
 import { convertResponsesMessages, convertResponsesTools, processResponsesStream } from "./openai-responses-shared.ts";
@@ -25,6 +25,21 @@ import { buildBaseOptions } from "./simple-options.ts";

 const OPENAI_TOOL_CALL_PROVIDERS = new Set(["openai", "openai-codex", "opencode"]);

+function hasHeader(headers: ProviderHeaders | undefined, name: string): boolean {
+	if (!headers) return false;
+	const expected = name.toLowerCase();
+	for (const [key, value] of Object.entries(headers)) {
+		if (key.toLowerCase() === expected && value !== null && value.trim().length > 0) return true;
+	}
+	return false;
+}
+
+function getClientApiKey(provider: string, apiKey: string | undefined, headers: ProviderHeaders | undefined): string {
+	if (apiKey) return apiKey;
+	if (hasHeader(headers, "authorization") || hasHeader(headers, "cf-aig-authorization")) return "unused";
+	throw new Error(`No API key for provider: ${provider}`);
+}
+
 /**
 * Resolve cache retention preference.
 * Defaults to "short" and uses PI_CACHE_RETENTION for backward compatibility.
@@ -109,13 +124,10 @@ export const stream: StreamFunction<"openai-responses", OpenAIResponsesOptions>

 		try {
 			// Create OpenAI client
-			const apiKey = options?.apiKey;
-			if (!apiKey) {
-				throw new Error(`No API key for provider: ${model.provider}`);
-			}
+			const apiKey = getClientApiKey(model.provider, options?.apiKey, options?.headers);
 			const cacheRetention = resolveCacheRetention(options?.cacheRetention, options?.env);
 			const cacheSessionId = cacheRetention === "none" ? undefined : options?.sessionId;
-			const client = createClient(model, context, apiKey, options?.headers, cacheSessionId, options?.env);
+			const client = createClient(model, context, apiKey, options?.headers, cacheSessionId);
 			let params = buildParams(model, context, options);
 			const nextParams = await options?.onPayload?.(params, model);
 			if (nextParams !== undefined) {
@@ -166,12 +178,9 @@ export const streamSimple: StreamFunction<"openai-responses", SimpleStreamOption
 	context: Context,
 	options?: SimpleStreamOptions,
 ): AssistantMessageEventStream => {
-	const apiKey = options?.apiKey;
-	if (!apiKey) {
-		throw new Error(`No API key for provider: ${model.provider}`);
-	}
+	getClientApiKey(model.provider, options?.apiKey, options?.headers);

-	const base = buildBaseOptions(model, options, apiKey);
+	const base = buildBaseOptions(model, options, options?.apiKey);
 	const clampedReasoning = options?.reasoning ? clampThinkingLevel(model, options.reasoning) : undefined;
 	const reasoningEffort = clampedReasoning === "off" ? undefined : clampedReasoning;

@@ -185,12 +194,11 @@ function createClient(
 	model: Model<"openai-responses">,
 	context: Context,
 	apiKey: string,
-	optionsHeaders?: Record<string, string>,
+	optionsHeaders?: ProviderHeaders,
 	sessionId?: string,
-	env?: ProviderEnv,
 ) {
 	const compat = getCompat(model);
-	const headers = { ...model.headers };
+	const headers: ProviderHeaders = { ...model.headers };
 	if (model.provider === "github-copilot") {
 		const hasImages = hasCopilotVisionInput(context.messages);
 		const copilotHeaders = buildCopilotDynamicHeaders({
@@ -212,20 +220,11 @@ function createClient(
 		Object.assign(headers, optionsHeaders);
 	}

-	const defaultHeaders =
-		model.provider === "cloudflare-ai-gateway"
-			? {
-					...headers,
-					Authorization: headers.Authorization ?? null,
-					"cf-aig-authorization": `Bearer ${apiKey}`,
-				}
-			: headers;
-
 	return new OpenAI({
 		apiKey,
-		baseURL: isCloudflareProvider(model.provider) ? resolveCloudflareBaseUrl(model, env) : model.baseUrl,
+		baseURL: model.baseUrl,
 		dangerouslyAllowBrowser: true,
-		defaultHeaders,
+		defaultHeaders: headers,
 	});
 }

@@ -13,9 +13,10 @@ import type {
 	ImagesFunction,
 	ImagesModel,
 	ImagesOptions,
+	ProviderHeaders,
 	TextContent,
 } from "../types.ts";
-import { headersToRecord } from "../utils/headers.ts";
+import { headersToRecord, providerHeadersToRecord } from "../utils/headers.ts";
 import { sanitizeSurrogates } from "../utils/sanitize-unicode.ts";

 interface OpenRouterGeneratedImage {
@@ -106,16 +107,13 @@ export const generateImages: ImagesFunction<"openrouter-images", ImagesOptions>
 function createClient(
 	model: ImagesModel<"openrouter-images">,
 	apiKey: string,
-	optionsHeaders?: Record<string, string>,
+	optionsHeaders?: ProviderHeaders,
 ): OpenAI {
 	return new OpenAI({
 		apiKey,
 		baseURL: model.baseUrl,
 		dangerouslyAllowBrowser: true,
-		defaultHeaders: {
-			...model.headers,
-			...optionsHeaders,
-		},
+		defaultHeaders: providerHeadersToRecord({ ...model.headers, ...optionsHeaders }),
 	});
 }

@@ -1,4 +1,4 @@
-import type { Api, ImagesApi, ImagesModel, Model, ProviderEnv } from "../types.ts";
+import type { Api, ImagesApi, ImagesModel, Model, ProviderEnv, ProviderHeaders } from "../types.ts";
 import type { OAuthCredentials } from "../utils/oauth/types.ts";

 /**
@@ -7,7 +7,7 @@ import type { OAuthCredentials } from "../utils/oauth/types.ts";
 */
 export interface ModelAuth {
 	apiKey?: string;
-	headers?: Record<string, string>;
+	headers?: ProviderHeaders;
 	baseUrl?: string;
 }

@@ -19,7 +19,6 @@ export * from "./api/mistral-conversations.lazy.ts";
 export * from "./api/openai-codex-responses.lazy.ts";
 export * from "./api/openai-completions.lazy.ts";
 export * from "./api/openai-responses.lazy.ts";
-export * from "./api-registry.ts";
 export * from "./env-api-keys.ts";
 export * from "./image-models.ts";
 export * from "./images.ts";
@@ -36,11 +35,12 @@ import { mistralConversationsApi } from "./api/mistral-conversations.lazy.ts";
 import { openAICodexResponsesApi } from "./api/openai-codex-responses.lazy.ts";
 import { openAICompletionsApi } from "./api/openai-completions.lazy.ts";
 import { openAIResponsesApi } from "./api/openai-responses.lazy.ts";
-import { clearApiProviders, getApiProvider, registerApiProvider } from "./api-registry.ts";
 import { getEnvApiKey } from "./env-api-keys.ts";
-import { getBuiltinModel, getBuiltinModels, getBuiltinProviders } from "./providers/all.ts";
+import { builtinModels, getBuiltinModel, getBuiltinModels, getBuiltinProviders } from "./providers/all.ts";
+import { createFauxCore, type FauxProviderRegistration, type RegisterFauxProviderOptions } from "./providers/faux.ts";
 import type {
 	Api,
+	ApiStreamOptions,
 	AssistantMessage,
 	AssistantMessageEventStream,
 	Context,
@@ -48,6 +48,7 @@ import type {
 	ProviderStreamOptions,
 	ProviderStreams,
 	SimpleStreamOptions,
+	StreamFunction,
 	StreamOptions,
 } from "./types.ts";

@@ -60,6 +61,113 @@ export const getModels = getBuiltinModels;
 /** @deprecated Static catalog read. Use `getBuiltinProviders` from "@earendil-works/pi-ai/providers/all" or `Models.getProviders()`. */
 export const getProviders = getBuiltinProviders;

+export type ApiStreamFunction = (
+	model: Model<Api>,
+	context: Context,
+	options?: StreamOptions,
+) => AssistantMessageEventStream;
+
+export type ApiStreamSimpleFunction = (
+	model: Model<Api>,
+	context: Context,
+	options?: SimpleStreamOptions,
+) => AssistantMessageEventStream;
+
+export interface ApiProvider<TApi extends Api = Api, TOptions extends StreamOptions = StreamOptions> {
+	api: TApi;
+	stream: StreamFunction<TApi, TOptions>;
+	streamSimple: StreamFunction<TApi, SimpleStreamOptions>;
+}
+
+interface ApiProviderInternal {
+	api: Api;
+	stream: ApiStreamFunction;
+	streamSimple: ApiStreamSimpleFunction;
+}
+
+type RegisteredApiProvider = {
+	provider: ApiProviderInternal;
+	sourceId?: string;
+};
+
+const apiProviderRegistry = new Map<string, RegisteredApiProvider>();
+
+function wrapStream<TApi extends Api, TOptions extends StreamOptions>(
+	api: TApi,
+	stream: StreamFunction<TApi, TOptions>,
+): ApiStreamFunction {
+	return (model, context, options) => {
+		if (model.api !== api) {
+			throw new Error(`Mismatched api: ${model.api} expected ${api}`);
+		}
+		return stream(model as Model<TApi>, context, options as TOptions);
+	};
+}
+
+function wrapStreamSimple<TApi extends Api>(
+	api: TApi,
+	streamSimple: StreamFunction<TApi, SimpleStreamOptions>,
+): ApiStreamSimpleFunction {
+	return (model, context, options) => {
+		if (model.api !== api) {
+			throw new Error(`Mismatched api: ${model.api} expected ${api}`);
+		}
+		return streamSimple(model as Model<TApi>, context, options);
+	};
+}
+
+export function registerApiProvider<TApi extends Api, TOptions extends StreamOptions>(
+	provider: ApiProvider<TApi, TOptions>,
+	sourceId?: string,
+): void {
+	apiProviderRegistry.set(provider.api, {
+		provider: {
+			api: provider.api,
+			stream: wrapStream(provider.api, provider.stream),
+			streamSimple: wrapStreamSimple(provider.api, provider.streamSimple),
+		},
+		sourceId,
+	});
+}
+
+export function getApiProvider(api: Api): ApiProviderInternal | undefined {
+	return apiProviderRegistry.get(api)?.provider;
+}
+
+export function getApiProviders(): ApiProviderInternal[] {
+	return Array.from(apiProviderRegistry.values(), (entry) => entry.provider);
+}
+
+export function unregisterApiProviders(sourceId: string): void {
+	for (const [api, entry] of apiProviderRegistry.entries()) {
+		if (entry.sourceId === sourceId) {
+			apiProviderRegistry.delete(api);
+		}
+	}
+}
+
+function clearApiProviders(): void {
+	apiProviderRegistry.clear();
+}
+
+export function registerFauxProvider(options: RegisterFauxProviderOptions = {}): FauxProviderRegistration {
+	const core = createFauxCore(options);
+	const sourceId = `faux-provider-${Math.random().toString(36).slice(2, 10)}`;
+	registerApiProvider({ api: core.api, stream: core.stream, streamSimple: core.streamSimple }, sourceId);
+	return {
+		api: core.api,
+		models: core.models,
+		getModel: core.getModel,
+		state: core.state,
+		setResponses: core.setResponses,
+		appendResponses: core.appendResponses,
+		getPendingResponseCount: core.getPendingResponseCount,
+		unregister() {
+			unregisterApiProviders(sourceId);
+		},
+	};
+}
+
 const BUILTIN_APIS: [Api, ProviderStreams][] = [
 	["anthropic-messages", anthropicMessagesApi()],
 	["openai-completions", openAICompletionsApi()],
@@ -72,6 +180,8 @@ const BUILTIN_APIS: [Api, ProviderStreams][] = [
 	["bedrock-converse-stream", bedrockConverseStreamApi()],
 ];

+const builtinApiProviderInstances = new Map<Api, ReturnType<typeof getApiProvider>>();
+
 /**
 * Registers the builtin API implementations into the api-registry without
 * clobbering existing entries: compat may load after a test or extension has
@@ -79,18 +189,23 @@ const BUILTIN_APIS: [Api, ProviderStreams][] = [
 */
 export function registerBuiltInApiProviders(): void {
 	for (const [api, streams] of BUILTIN_APIS) {
-		if (getApiProvider(api)) continue;
-		registerApiProvider({ api, stream: streams.stream, streamSimple: streams.streamSimple });
+		if (!getApiProvider(api)) {
+			registerApiProvider({ api, stream: streams.stream, streamSimple: streams.streamSimple });
+		}
+		builtinApiProviderInstances.set(api, getApiProvider(api));
 	}
 }

 export function resetApiProviders(): void {
 	clearApiProviders();
+	builtinApiProviderInstances.clear();
 	registerBuiltInApiProviders();
 }

 registerBuiltInApiProviders();

+const compatModels = builtinModels();
+
 function hasExplicitApiKey(apiKey: string | undefined): apiKey is string {
 	return typeof apiKey === "string" && apiKey.trim().length > 0;
 }
@@ -105,6 +220,11 @@ function withEnvApiKey<TOptions extends StreamOptions>(
 	return { ...options, apiKey } as TOptions;
 }

+function shouldUseBuiltinModels(model: Model<Api>): boolean {
+	const builtin = compatModels.getModel(model.provider, model.id);
+	return builtin?.api === model.api && getApiProvider(model.api) === builtinApiProviderInstances.get(model.api);
+}
+
 function resolveApiProvider(api: Api) {
 	const provider = getApiProvider(api);
 	if (!provider) {
@@ -118,6 +238,9 @@ export function stream<TApi extends Api>(
 	context: Context,
 	options?: ProviderStreamOptions,
 ): AssistantMessageEventStream {
+	if (shouldUseBuiltinModels(model)) {
+		return compatModels.stream(model, context, options as ApiStreamOptions<TApi> | undefined);
+	}
 	const provider = resolveApiProvider(model.api);
 	return provider.stream(model, context, withEnvApiKey(model, options) as StreamOptions);
 }
@@ -136,6 +259,9 @@ export function streamSimple<TApi extends Api>(
 	context: Context,
 	options?: SimpleStreamOptions,
 ): AssistantMessageEventStream {
+	if (shouldUseBuiltinModels(model)) {
+		return compatModels.streamSimple(model, context, options);
+	}
 	const provider = resolveApiProvider(model.api);
 	return provider.streamSimple(model, context, withEnvApiKey(model, options));
 }
@@ -11,6 +11,7 @@ import type {
 	Context,
 	Model,
 	ModelThinkingLevel,
+	ProviderHeaders,
 	ProviderStreams,
 	SimpleStreamOptions,
 	StreamOptions,
@@ -33,7 +34,7 @@ export interface Provider<TApi extends Api = Api> {
 	readonly name: string;

 	readonly baseUrl?: string;
-	readonly headers?: Record<string, string>;
+	readonly headers?: ProviderHeaders;

 	/**
 	 * Required: at least one of `apiKey`/`oauth`. Every provider has auth
@@ -287,7 +288,7 @@ export interface CreateProviderOptions<TApi extends Api = Api> {
 	/** Display name. Default: `id`. */
 	name?: string;
 	baseUrl?: string;
-	headers?: Record<string, string>;
+	headers?: ProviderHeaders;
 	/** Required — every provider has auth semantics, even ambient/keyless ones. */
 	auth: ProviderAuth;
 	/** Initial model list (empty for purely dynamic providers). */
@@ -10,7 +10,7 @@ export const ANT_LING_MODELS = {
 		api: "openai-completions",
 		provider: "ant-ling",
 		baseUrl: "https://api.ant-ling.com/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"ant-ling","supportsLongCacheRetention":false},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -28,7 +28,7 @@ export const ANT_LING_MODELS = {
 		api: "openai-completions",
 		provider: "ant-ling",
 		baseUrl: "https://api.ant-ling.com/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"ant-ling","supportsLongCacheRetention":false},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -46,7 +46,7 @@ export const ANT_LING_MODELS = {
 		api: "openai-completions",
 		provider: "ant-ling",
 		baseUrl: "https://api.ant-ling.com/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsLongCacheRetention":false,"thinkingFormat":"ant-ling"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"ant-ling","supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"off":null,"minimal":null,"low":null,"medium":null,"high":"high","xhigh":"xhigh"},
 		input: ["text"],
@@ -10,6 +10,7 @@ export const CEREBRAS_MODELS = {
 		api: "openai-completions",
 		provider: "cerebras",
 		baseUrl: "https://api.cerebras.ai/v1",
+		compat: {"supportsStore":false,"supportsDeveloperRole":false},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -27,6 +28,7 @@ export const CEREBRAS_MODELS = {
 		api: "openai-completions",
 		provider: "cerebras",
 		baseUrl: "https://api.cerebras.ai/v1",
+		compat: {"supportsStore":false,"supportsDeveloperRole":false},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -10,6 +10,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: false,
 		input: ["text", "image"],
 		cost: {
@@ -27,6 +28,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: false,
 		input: ["text", "image"],
 		cost: {
@@ -44,6 +46,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: false,
 		input: ["text", "image"],
 		cost: {
@@ -61,6 +64,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: false,
 		input: ["text", "image"],
 		cost: {
@@ -78,6 +82,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: false,
 		input: ["text", "image"],
 		cost: {
@@ -95,6 +100,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: false,
 		input: ["text", "image"],
 		cost: {
@@ -112,7 +118,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
-		compat: {"forceAdaptiveThinking":true},
+		compat: {"sendSessionAffinityHeaders":true,"forceAdaptiveThinking":true},
 		reasoning: true,
 		thinkingLevelMap: {"off":null,"xhigh":"xhigh"},
 		input: ["text", "image"],
@@ -131,6 +137,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -148,6 +155,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -165,6 +173,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -182,6 +191,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -199,7 +209,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
-		compat: {"forceAdaptiveThinking":true},
+		compat: {"sendSessionAffinityHeaders":true,"forceAdaptiveThinking":true},
 		reasoning: true,
 		thinkingLevelMap: {"xhigh":"max"},
 		input: ["text", "image"],
@@ -218,7 +228,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
-		compat: {"forceAdaptiveThinking":true,"supportsTemperature":false},
+		compat: {"sendSessionAffinityHeaders":true,"forceAdaptiveThinking":true,"supportsTemperature":false},
 		reasoning: true,
 		thinkingLevelMap: {"xhigh":"xhigh"},
 		input: ["text", "image"],
@@ -237,7 +247,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
-		compat: {"forceAdaptiveThinking":true,"supportsTemperature":false},
+		compat: {"sendSessionAffinityHeaders":true,"forceAdaptiveThinking":true,"supportsTemperature":false},
 		reasoning: true,
 		thinkingLevelMap: {"xhigh":"xhigh"},
 		input: ["text", "image"],
@@ -256,6 +266,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -273,6 +284,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
+		compat: {"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -290,7 +302,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "anthropic-messages",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/anthropic",
-		compat: {"forceAdaptiveThinking":true},
+		compat: {"sendSessionAffinityHeaders":true,"forceAdaptiveThinking":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -587,7 +599,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/compat",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -605,7 +617,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/compat",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -623,7 +635,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/compat",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -641,7 +653,7 @@ export const CLOUDFLARE_AI_GATEWAY_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-ai-gateway",
 		baseUrl: "https://gateway.ai.cloudflare.com/v1/{CLOUDFLARE_ACCOUNT_ID}/{CLOUDFLARE_GATEWAY_ID}/compat",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -1,9 +1,9 @@
 import { anthropicMessagesApi } from "../api/anthropic-messages.lazy.ts";
 import { openAICompletionsApi } from "../api/openai-completions.lazy.ts";
 import { openAIResponsesApi } from "../api/openai-responses.lazy.ts";
-import { envApiKeyAuth } from "../auth/helpers.ts";
 import { createProvider, type Provider } from "../models.ts";
 import { CLOUDFLARE_AI_GATEWAY_MODELS } from "./cloudflare-ai-gateway.models.ts";
+import { cloudflareAIGatewayAuth } from "./cloudflare-auth.ts";

 export function cloudflareAIGatewayProvider(): Provider<
 	"anthropic-messages" | "openai-completions" | "openai-responses"
@@ -11,7 +11,7 @@ export function cloudflareAIGatewayProvider(): Provider<
 	return createProvider({
 		id: "cloudflare-ai-gateway",
 		name: "Cloudflare AI Gateway",
-		auth: { apiKey: envApiKeyAuth("Cloudflare API key", ["CLOUDFLARE_API_KEY"]) },
+		auth: { apiKey: cloudflareAIGatewayAuth() },
 		models: Object.values(CLOUDFLARE_AI_GATEWAY_MODELS),
 		api: {
 			"anthropic-messages": anthropicMessagesApi(),
@@ -0,0 +1,108 @@
+import type { ApiKeyAuth, ApiKeyCredential, AuthContext } from "../auth/types.ts";
+import type { Api, ImagesApi, ImagesModel, Model, ProviderEnv } from "../types.ts";
+
+const CLOUDFLARE_API_KEY = "CLOUDFLARE_API_KEY";
+const CLOUDFLARE_ACCOUNT_ID = "CLOUDFLARE_ACCOUNT_ID";
+const CLOUDFLARE_GATEWAY_ID = "CLOUDFLARE_GATEWAY_ID";
+
+type CloudflareAuthKind = "workers-ai" | "ai-gateway";
+
+async function resolveValue(input: {
+	name: string;
+	ctx: AuthContext;
+	credential: ApiKeyCredential | undefined;
+}): Promise<string | undefined> {
+	if (input.credential) {
+		if (input.name === CLOUDFLARE_API_KEY) return input.credential.key;
+		return input.credential.metadata?.[input.name];
+	}
+	return input.ctx.env(input.name);
+}
+
+function resolveCloudflareBaseUrl(
+	model: Model<Api> | ImagesModel<ImagesApi>,
+	accountId: string,
+	gatewayId: string | undefined,
+): string {
+	return model.baseUrl
+		.replaceAll(`{${CLOUDFLARE_ACCOUNT_ID}}`, accountId)
+		.replaceAll(`{${CLOUDFLARE_GATEWAY_ID}}`, gatewayId ?? "");
+}
+
+async function resolveCloudflareEnv(input: {
+	kind: CloudflareAuthKind;
+	model: Model<Api> | ImagesModel<ImagesApi>;
+	ctx: AuthContext;
+	credential: ApiKeyCredential | undefined;
+}): Promise<{ apiKey: string; env: ProviderEnv; baseUrl: string; source: string } | undefined> {
+	const apiKey = await resolveValue({ name: CLOUDFLARE_API_KEY, ctx: input.ctx, credential: input.credential });
+	const accountId = await resolveValue({ name: CLOUDFLARE_ACCOUNT_ID, ctx: input.ctx, credential: input.credential });
+	const gatewayId =
+		input.kind === "ai-gateway"
+			? await resolveValue({ name: CLOUDFLARE_GATEWAY_ID, ctx: input.ctx, credential: input.credential })
+			: undefined;
+
+	if (!apiKey || !accountId || (input.kind === "ai-gateway" && !gatewayId)) return undefined;
+
+	return {
+		apiKey,
+		env: {
+			CLOUDFLARE_ACCOUNT_ID: accountId,
+			...(gatewayId ? { CLOUDFLARE_GATEWAY_ID: gatewayId } : {}),
+		},
+		baseUrl: resolveCloudflareBaseUrl(input.model, accountId, gatewayId),
+		source: input.credential ? "stored credential" : CLOUDFLARE_API_KEY,
+	};
+}
+
+export function cloudflareWorkersAIAuth(): ApiKeyAuth {
+	return {
+		name: "Cloudflare API key",
+		login: async (callbacks) => {
+			const key = await callbacks.prompt({ type: "secret", message: "Enter Cloudflare API key" });
+			const accountId = await callbacks.prompt({ type: "text", message: "Enter Cloudflare account ID" });
+			return { type: "api-key", key, metadata: { CLOUDFLARE_ACCOUNT_ID: accountId } };
+		},
+		resolve: async ({ model, ctx, credential }) => {
+			const resolved = await resolveCloudflareEnv({ kind: "workers-ai", model, ctx, credential });
+			if (!resolved) return undefined;
+			return {
+				auth: { apiKey: resolved.apiKey, baseUrl: resolved.baseUrl },
+				env: resolved.env,
+				source: resolved.source,
+			};
+		},
+	};
+}
+
+export function cloudflareAIGatewayAuth(): ApiKeyAuth {
+	return {
+		name: "Cloudflare API key",
+		login: async (callbacks) => {
+			const key = await callbacks.prompt({ type: "secret", message: "Enter Cloudflare API key" });
+			const accountId = await callbacks.prompt({ type: "text", message: "Enter Cloudflare account ID" });
+			const gatewayId = await callbacks.prompt({ type: "text", message: "Enter Cloudflare AI Gateway ID" });
+			return {
+				type: "api-key",
+				key,
+				metadata: { CLOUDFLARE_ACCOUNT_ID: accountId, CLOUDFLARE_GATEWAY_ID: gatewayId },
+			};
+		},
+		resolve: async ({ model, ctx, credential }) => {
+			const resolved = await resolveCloudflareEnv({ kind: "ai-gateway", model, ctx, credential });
+			if (!resolved) return undefined;
+			return {
+				auth: {
+					headers: {
+						"cf-aig-authorization": `Bearer ${resolved.apiKey}`,
+						Authorization: null,
+						"x-api-key": null,
+					},
+					baseUrl: resolved.baseUrl,
+				},
+				env: resolved.env,
+				source: resolved.source,
+			};
+		},
+	};
+}
@@ -10,7 +10,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -28,7 +28,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -46,7 +46,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -64,7 +64,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: false,
 		input: ["text", "image"],
 		cost: {
@@ -82,7 +82,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -100,7 +100,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -118,7 +118,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -136,7 +136,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -154,7 +154,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -172,7 +172,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -190,7 +190,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -208,7 +208,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -226,7 +226,7 @@ export const CLOUDFLARE_WORKERS_AI_MODELS = {
 		api: "openai-completions",
 		provider: "cloudflare-workers-ai",
 		baseUrl: "https://api.cloudflare.com/client/v4/accounts/{CLOUDFLARE_ACCOUNT_ID}/ai/v1",
-		compat: {"sendSessionAffinityHeaders":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsLongCacheRetention":false,"sendSessionAffinityHeaders":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -1,13 +1,13 @@
 import { openAICompletionsApi } from "../api/openai-completions.lazy.ts";
-import { envApiKeyAuth } from "../auth/helpers.ts";
 import { createProvider, type Provider } from "../models.ts";
+import { cloudflareWorkersAIAuth } from "./cloudflare-auth.ts";
 import { CLOUDFLARE_WORKERS_AI_MODELS } from "./cloudflare-workers-ai.models.ts";

 export function cloudflareWorkersAIProvider(): Provider<"openai-completions"> {
 	return createProvider({
 		id: "cloudflare-workers-ai",
 		name: "Cloudflare Workers AI",
-		auth: { apiKey: envApiKeyAuth("Cloudflare API key", ["CLOUDFLARE_API_KEY"]) },
+		auth: { apiKey: cloudflareWorkersAIAuth() },
 		models: Object.values(CLOUDFLARE_WORKERS_AI_MODELS),
 		api: openAICompletionsApi(),
 	});
@@ -10,7 +10,7 @@ export const DEEPSEEK_MODELS = {
 		api: "openai-completions",
 		provider: "deepseek",
 		baseUrl: "https://api.deepseek.com",
-		compat: {"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null,"high":"high","xhigh":"max"},
 		input: ["text"],
@@ -29,7 +29,7 @@ export const DEEPSEEK_MODELS = {
 		api: "openai-completions",
 		provider: "deepseek",
 		baseUrl: "https://api.deepseek.com",
-		compat: {"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null,"high":"high","xhigh":"max"},
 		input: ["text"],
@@ -1,4 +1,3 @@
-import { registerApiProvider, unregisterApiProviders } from "../api-registry.ts";
 import { createProvider, type Provider } from "../models.ts";
 import type {
 	AssistantMessage,
@@ -401,7 +400,7 @@ async function streamWithDeltas(
 	stream.end(message);
 }

-function createFauxCore(options: RegisterFauxProviderOptions) {
+export function createFauxCore(options: RegisterFauxProviderOptions) {
 	const api = options.api ?? randomId(DEFAULT_API);
 	const provider = options.provider ?? DEFAULT_PROVIDER;
 	const minTokenSize = Math.max(
@@ -508,25 +507,6 @@ function createFauxCore(options: RegisterFauxProviderOptions) {
 	};
 }

-/** Registers the faux api into the legacy global api-registry. */
-export function registerFauxProvider(options: RegisterFauxProviderOptions = {}): FauxProviderRegistration {
-	const core = createFauxCore(options);
-	const sourceId = randomId("faux-provider");
-	registerApiProvider({ api: core.api, stream: core.stream, streamSimple: core.streamSimple }, sourceId);
-	return {
-		api: core.api,
-		models: core.models,
-		getModel: core.getModel,
-		state: core.state,
-		setResponses: core.setResponses,
-		appendResponses: core.appendResponses,
-		getPendingResponseCount: core.getPendingResponseCount,
-		unregister() {
-			unregisterApiProviders(sourceId);
-		},
-	};
-}
-
 /**
 * Faux provider for tests built on explicit `Models` collections:
 *
@@ -4,6 +4,24 @@
 import type { Model } from "../types.ts";

 export const HUGGINGFACE_MODELS = {
+	"MiniMaxAI/MiniMax-M2": {
+		id: "MiniMaxAI/MiniMax-M2",
+		name: "MiniMax-M2",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text"],
+		cost: {
+			input: 0.3,
+			output: 1.2,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 204800,
+		maxTokens: 128000,
+	} satisfies Model<"openai-completions">,
 	"MiniMaxAI/MiniMax-M2.1": {
 		id: "MiniMaxAI/MiniMax-M2.1",
 		name: "MiniMax-M2.1",
@@ -58,6 +76,42 @@ export const HUGGINGFACE_MODELS = {
 		contextWindow: 204800,
 		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
+	"MiniMaxAI/MiniMax-M3": {
+		id: "MiniMaxAI/MiniMax-M3",
+		name: "MiniMax-M3",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0.3,
+			output: 1.2,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 524288,
+		maxTokens: 128000,
+	} satisfies Model<"openai-completions">,
+	"Qwen/Qwen3-235B-A22B": {
+		id: "Qwen/Qwen3-235B-A22B",
+		name: "Qwen3 235B-A22B",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text"],
+		cost: {
+			input: 0.2,
+			output: 0.8,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 40960,
+		maxTokens: 16384,
+	} satisfies Model<"openai-completions">,
 	"Qwen/Qwen3-235B-A22B-Thinking-2507": {
 		id: "Qwen/Qwen3-235B-A22B-Thinking-2507",
 		name: "Qwen3-235B-A22B-Thinking-2507",
@@ -76,6 +130,42 @@ export const HUGGINGFACE_MODELS = {
 		contextWindow: 262144,
 		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
+	"Qwen/Qwen3-32B": {
+		id: "Qwen/Qwen3-32B",
+		name: "Qwen3 32B",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text"],
+		cost: {
+			input: 0.29,
+			output: 0.59,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 131072,
+		maxTokens: 16384,
+	} satisfies Model<"openai-completions">,
+	"Qwen/Qwen3-Coder-30B-A3B-Instruct": {
+		id: "Qwen/Qwen3-Coder-30B-A3B-Instruct",
+		name: "Qwen3-Coder 30B-A3B Instruct",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: false,
+		input: ["text"],
+		cost: {
+			input: 0.07,
+			output: 0.26,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 65536,
+	} satisfies Model<"openai-completions">,
 	"Qwen/Qwen3-Coder-480B-A35B-Instruct": {
 		id: "Qwen/Qwen3-Coder-480B-A35B-Instruct",
 		name: "Qwen3-Coder-480B-A35B-Instruct",
@@ -148,6 +238,60 @@ export const HUGGINGFACE_MODELS = {
 		contextWindow: 262144,
 		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
+	"Qwen/Qwen3.5-122B-A10B": {
+		id: "Qwen/Qwen3.5-122B-A10B",
+		name: "Qwen3.5 122B-A10B",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0.4,
+			output: 3.2,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 65536,
+	} satisfies Model<"openai-completions">,
+	"Qwen/Qwen3.5-27B": {
+		id: "Qwen/Qwen3.5-27B",
+		name: "Qwen3.5 27B",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0.3,
+			output: 2.4,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 65536,
+	} satisfies Model<"openai-completions">,
+	"Qwen/Qwen3.5-35B-A3B": {
+		id: "Qwen/Qwen3.5-35B-A3B",
+		name: "Qwen3.5 35B-A3B",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0.25,
+			output: 2,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 65536,
+	} satisfies Model<"openai-completions">,
 	"Qwen/Qwen3.5-397B-A17B": {
 		id: "Qwen/Qwen3.5-397B-A17B",
 		name: "Qwen3.5-397B-A17B",
@@ -166,6 +310,42 @@ export const HUGGINGFACE_MODELS = {
 		contextWindow: 262144,
 		maxTokens: 32768,
 	} satisfies Model<"openai-completions">,
+	"Qwen/Qwen3.5-9B": {
+		id: "Qwen/Qwen3.5-9B",
+		name: "Qwen3.5 9B",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0.17,
+			output: 0.25,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 65536,
+	} satisfies Model<"openai-completions">,
+	"Qwen/Qwen3.6-35B-A3B": {
+		id: "Qwen/Qwen3.6-35B-A3B",
+		name: "Qwen3.6 35B-A3B",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0.15,
+			output: 0.95,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 65536,
+	} satisfies Model<"openai-completions">,
 	"XiaomiMiMo/MiMo-V2-Flash": {
 		id: "XiaomiMiMo/MiMo-V2-Flash",
 		name: "MiMo-V2-Flash",
@@ -184,6 +364,24 @@ export const HUGGINGFACE_MODELS = {
 		contextWindow: 262144,
 		maxTokens: 4096,
 	} satisfies Model<"openai-completions">,
+	"deepseek-ai/DeepSeek-R1": {
+		id: "deepseek-ai/DeepSeek-R1",
+		name: "DeepSeek-R1",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text"],
+		cost: {
+			input: 0.7,
+			output: 2.5,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 64000,
+		maxTokens: 32768,
+	} satisfies Model<"openai-completions">,
 	"deepseek-ai/DeepSeek-R1-0528": {
 		id: "deepseek-ai/DeepSeek-R1-0528",
 		name: "DeepSeek-R1-0528",
@@ -220,6 +418,24 @@ export const HUGGINGFACE_MODELS = {
 		contextWindow: 163840,
 		maxTokens: 65536,
 	} satisfies Model<"openai-completions">,
+	"deepseek-ai/DeepSeek-V4-Flash": {
+		id: "deepseek-ai/DeepSeek-V4-Flash",
+		name: "DeepSeek V4 Flash",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text"],
+		cost: {
+			input: 0.14,
+			output: 0.28,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 1048576,
+		maxTokens: 384000,
+	} satisfies Model<"openai-completions">,
 	"deepseek-ai/DeepSeek-V4-Pro": {
 		id: "deepseek-ai/DeepSeek-V4-Pro",
 		name: "DeepSeek V4 Pro",
@@ -238,6 +454,60 @@ export const HUGGINGFACE_MODELS = {
 		contextWindow: 1048576,
 		maxTokens: 393216,
 	} satisfies Model<"openai-completions">,
+	"google/gemma-4-26B-A4B-it": {
+		id: "google/gemma-4-26B-A4B-it",
+		name: "Gemma 4 26B A4B IT",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0.13,
+			output: 0.4,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 32768,
+	} satisfies Model<"openai-completions">,
+	"google/gemma-4-31B-it": {
+		id: "google/gemma-4-31B-it",
+		name: "Gemma 4 31B IT",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0.14,
+			output: 0.4,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 32768,
+	} satisfies Model<"openai-completions">,
+	"meta-llama/Llama-3.3-70B-Instruct": {
+		id: "meta-llama/Llama-3.3-70B-Instruct",
+		name: "Llama-3.3-70B-Instruct",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: false,
+		input: ["text"],
+		cost: {
+			input: 0.59,
+			output: 0.79,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 131072,
+		maxTokens: 4096,
+	} satisfies Model<"openai-completions">,
 	"moonshotai/Kimi-K2-Instruct": {
 		id: "moonshotai/Kimi-K2-Instruct",
 		name: "Kimi-K2-Instruct",
@@ -328,6 +598,114 @@ export const HUGGINGFACE_MODELS = {
 		contextWindow: 262144,
 		maxTokens: 262144,
 	} satisfies Model<"openai-completions">,
+	"moonshotai/Kimi-K2.7-Code": {
+		id: "moonshotai/Kimi-K2.7-Code",
+		name: "Kimi K2.7 Code",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0.95,
+			output: 4,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 262144,
+	} satisfies Model<"openai-completions">,
+	"stepfun-ai/Step-3.5-Flash": {
+		id: "stepfun-ai/Step-3.5-Flash",
+		name: "Step 3.5 Flash",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text"],
+		cost: {
+			input: 0.1,
+			output: 0.3,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 256000,
+	} satisfies Model<"openai-completions">,
+	"zai-org/GLM-4.5": {
+		id: "zai-org/GLM-4.5",
+		name: "GLM-4.5",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text"],
+		cost: {
+			input: 0.6,
+			output: 2.2,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 131072,
+		maxTokens: 98304,
+	} satisfies Model<"openai-completions">,
+	"zai-org/GLM-4.5-Air": {
+		id: "zai-org/GLM-4.5-Air",
+		name: "GLM-4.5-Air",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text"],
+		cost: {
+			input: 0.13,
+			output: 0.85,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 131072,
+		maxTokens: 98304,
+	} satisfies Model<"openai-completions">,
+	"zai-org/GLM-4.5V": {
+		id: "zai-org/GLM-4.5V",
+		name: "GLM-4.5V",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text", "image"],
+		cost: {
+			input: 0.6,
+			output: 1.8,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 65536,
+		maxTokens: 16384,
+	} satisfies Model<"openai-completions">,
+	"zai-org/GLM-4.6": {
+		id: "zai-org/GLM-4.6",
+		name: "GLM-4.6",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text"],
+		cost: {
+			input: 0.55,
+			output: 2.2,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 204800,
+		maxTokens: 131072,
+	} satisfies Model<"openai-completions">,
 	"zai-org/GLM-4.7": {
 		id: "zai-org/GLM-4.7",
 		name: "GLM-4.7",
@@ -400,4 +778,22 @@ export const HUGGINGFACE_MODELS = {
 		contextWindow: 202752,
 		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
+	"zai-org/GLM-5.2": {
+		id: "zai-org/GLM-5.2",
+		name: "GLM-5.2",
+		api: "openai-completions",
+		provider: "huggingface",
+		baseUrl: "https://router.huggingface.co/v1",
+		compat: {"supportsDeveloperRole":false},
+		reasoning: true,
+		input: ["text"],
+		cost: {
+			input: 1.4,
+			output: 4.4,
+			cacheRead: 0,
+			cacheWrite: 0,
+		},
+		contextWindow: 262144,
+		maxTokens: 131072,
+	} satisfies Model<"openai-completions">,
 } as const;
@@ -10,7 +10,7 @@ export const OPENCODE_GO_MODELS = {
 		api: "openai-completions",
 		provider: "opencode-go",
 		baseUrl: "https://opencode.ai/zen/go/v1",
-		compat: {"maxTokensField":"max_tokens","requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens","requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null,"high":"high","xhigh":"max"},
 		input: ["text"],
@@ -29,7 +29,7 @@ export const OPENCODE_GO_MODELS = {
 		api: "openai-completions",
 		provider: "opencode-go",
 		baseUrl: "https://opencode.ai/zen/go/v1",
-		compat: {"maxTokensField":"max_tokens","requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens","requiresReasoningContentOnAssistantMessages":true,"thinkingFormat":"deepseek"},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null,"high":"high","xhigh":"max"},
 		input: ["text"],
@@ -48,7 +48,7 @@ export const OPENCODE_GO_MODELS = {
 		api: "openai-completions",
 		provider: "opencode-go",
 		baseUrl: "https://opencode.ai/zen/go/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -66,7 +66,7 @@ export const OPENCODE_GO_MODELS = {
 		api: "openai-completions",
 		provider: "opencode-go",
 		baseUrl: "https://opencode.ai/zen/go/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		thinkingLevelMap: {"off":null,"minimal":null,"low":null,"medium":null,"high":"high","xhigh":"max"},
 		input: ["text"],
@@ -85,7 +85,7 @@ export const OPENCODE_GO_MODELS = {
 		api: "openai-completions",
 		provider: "opencode-go",
 		baseUrl: "https://opencode.ai/zen/go/v1",
-		compat: {"thinkingFormat":"deepseek","supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"thinkingFormat":"deepseek","supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null},
 		input: ["text", "image"],
@@ -104,7 +104,7 @@ export const OPENCODE_GO_MODELS = {
 		api: "openai-completions",
 		provider: "opencode-go",
 		baseUrl: "https://opencode.ai/zen/go/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -122,7 +122,7 @@ export const OPENCODE_GO_MODELS = {
 		api: "openai-completions",
 		provider: "opencode-go",
 		baseUrl: "https://opencode.ai/zen/go/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -140,7 +140,7 @@ export const OPENCODE_GO_MODELS = {
 		api: "openai-completions",
 		provider: "opencode-go",
 		baseUrl: "https://opencode.ai/zen/go/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -158,7 +158,7 @@ export const OPENCODE_GO_MODELS = {
 		api: "openai-completions",
 		provider: "opencode-go",
 		baseUrl: "https://opencode.ai/zen/go/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -193,7 +193,7 @@ export const OPENCODE_GO_MODELS = {
 		api: "openai-completions",
 		provider: "opencode-go",
 		baseUrl: "https://opencode.ai/zen/go/v1",
-		compat: {"thinkingFormat":"qwen","maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"thinkingFormat":"qwen","maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -10,7 +10,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -188,7 +188,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens","supportsLongCacheRetention":false,"requiresReasoningContentOnAssistantMessages":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens","supportsLongCacheRetention":false,"requiresReasoningContentOnAssistantMessages":true},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null,"high":"high","xhigh":"max"},
 		input: ["text"],
@@ -207,7 +207,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens","requiresReasoningContentOnAssistantMessages":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens","requiresReasoningContentOnAssistantMessages":true},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null,"high":"high","xhigh":"max"},
 		input: ["text"],
@@ -226,7 +226,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens","supportsLongCacheRetention":false,"requiresReasoningContentOnAssistantMessages":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens","supportsLongCacheRetention":false,"requiresReasoningContentOnAssistantMessages":true},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null,"high":"high","xhigh":"max"},
 		input: ["text"],
@@ -299,7 +299,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -317,7 +317,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -329,6 +329,24 @@ export const OPENCODE_MODELS = {
 		contextWindow: 204800,
 		maxTokens: 131072,
 	} satisfies Model<"openai-completions">,
+	"glm-5.2": {
+		id: "glm-5.2",
+		name: "GLM-5.2",
+		api: "openai-completions",
+		provider: "opencode",
+		baseUrl: "https://opencode.ai/zen/v1",
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
+		reasoning: true,
+		input: ["text"],
+		cost: {
+			input: 1.4,
+			output: 4.4,
+			cacheRead: 0.26,
+			cacheWrite: 0,
+		},
+		contextWindow: 1000000,
+		maxTokens: 131072,
+	} satisfies Model<"openai-completions">,
 	"gpt-5": {
 		id: "gpt-5",
 		name: "GPT-5",
@@ -623,7 +641,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"supportsReasoningEffort":false,"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		thinkingLevelMap: {"off":null,"minimal":null,"low":null,"medium":null},
 		input: ["text", "image"],
@@ -642,7 +660,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens","supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens","supportsLongCacheRetention":false},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -660,7 +678,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"thinkingFormat":"deepseek","supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"thinkingFormat":"deepseek","supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsLongCacheRetention":false},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -678,7 +696,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -696,7 +714,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -714,7 +732,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens","supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens","supportsLongCacheRetention":false},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -732,7 +750,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -750,7 +768,7 @@ export const OPENCODE_MODELS = {
 		api: "openai-completions",
 		provider: "opencode",
 		baseUrl: "https://opencode.ai/zen/v1",
-		compat: {"maxTokensField":"max_tokens"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"maxTokensField":"max_tokens"},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -10,7 +10,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"off":null,"minimal":null,"low":null,"medium":null},
 		input: ["text"],
@@ -29,7 +29,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"together"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null},
 		input: ["text", "image"],
@@ -48,7 +48,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -66,7 +66,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -84,7 +84,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"together"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null},
 		input: ["text", "image"],
@@ -103,7 +103,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"together"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null},
 		input: ["text", "image"],
@@ -122,7 +122,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"together"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null},
 		input: ["text"],
@@ -141,7 +141,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -159,7 +159,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"together"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null,"high":"high","xhigh":null},
 		input: ["text"],
@@ -178,7 +178,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -196,7 +196,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"together"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null},
 		input: ["text", "image"],
@@ -215,7 +215,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -233,7 +233,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"together"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null},
 		input: ["text", "image"],
@@ -252,7 +252,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"together"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null},
 		input: ["text"],
@@ -271,7 +271,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"together"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null},
 		input: ["text"],
@@ -290,7 +290,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"openai"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"openai","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"off":null,"minimal":null},
 		input: ["text"],
@@ -309,7 +309,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"openai"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"maxTokensField":"max_tokens","thinkingFormat":"openai","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"off":null,"minimal":null},
 		input: ["text"],
@@ -328,7 +328,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"together"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null},
 		input: ["text"],
@@ -347,7 +347,7 @@ export const TOGETHER_MODELS = {
 		api: "openai-completions",
 		provider: "together",
 		baseUrl: "https://api.together.ai/v1",
-		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","supportsStrictMode":false,"supportsLongCacheRetention":false,"thinkingFormat":"together"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"maxTokensField":"max_tokens","thinkingFormat":"together","supportsStrictMode":false,"supportsLongCacheRetention":false},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":null,"medium":null},
 		input: ["text"],
@@ -10,6 +10,7 @@ export const XAI_MODELS = {
 		api: "openai-completions",
 		provider: "xai",
 		baseUrl: "https://api.x.ai/v1",
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -27,6 +28,7 @@ export const XAI_MODELS = {
 		api: "openai-completions",
 		provider: "xai",
 		baseUrl: "https://api.x.ai/v1",
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -44,6 +46,7 @@ export const XAI_MODELS = {
 		api: "openai-completions",
 		provider: "xai",
 		baseUrl: "https://api.x.ai/v1",
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false},
 		reasoning: false,
 		input: ["text", "image"],
 		cost: {
@@ -61,6 +64,7 @@ export const XAI_MODELS = {
 		api: "openai-completions",
 		provider: "xai",
 		baseUrl: "https://api.x.ai/v1",
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -78,6 +82,7 @@ export const XAI_MODELS = {
 		api: "openai-completions",
 		provider: "xai",
 		baseUrl: "https://api.x.ai/v1",
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -95,6 +100,7 @@ export const XAI_MODELS = {
 		api: "openai-completions",
 		provider: "xai",
 		baseUrl: "https://api.x.ai/v1",
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -112,6 +118,7 @@ export const XAI_MODELS = {
 		api: "openai-completions",
 		provider: "xai",
 		baseUrl: "https://api.x.ai/v1",
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false},
 		reasoning: false,
 		input: ["text"],
 		cost: {
@@ -10,7 +10,7 @@ export const ZAI_CODING_CN_MODELS = {
 		api: "openai-completions",
 		provider: "zai-coding-cn",
 		baseUrl: "https://open.bigmodel.cn/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"thinkingFormat":"zai"},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -28,7 +28,7 @@ export const ZAI_CODING_CN_MODELS = {
 		api: "openai-completions",
 		provider: "zai-coding-cn",
 		baseUrl: "https://open.bigmodel.cn/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai","zaiToolStream":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"thinkingFormat":"zai","zaiToolStream":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -46,7 +46,7 @@ export const ZAI_CODING_CN_MODELS = {
 		api: "openai-completions",
 		provider: "zai-coding-cn",
 		baseUrl: "https://open.bigmodel.cn/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai","zaiToolStream":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"thinkingFormat":"zai","zaiToolStream":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -64,7 +64,7 @@ export const ZAI_CODING_CN_MODELS = {
 		api: "openai-completions",
 		provider: "zai-coding-cn",
 		baseUrl: "https://open.bigmodel.cn/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai","zaiToolStream":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"thinkingFormat":"zai","zaiToolStream":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -82,7 +82,7 @@ export const ZAI_CODING_CN_MODELS = {
 		api: "openai-completions",
 		provider: "zai-coding-cn",
 		baseUrl: "https://open.bigmodel.cn/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai","supportsReasoningEffort":true,"zaiToolStream":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"thinkingFormat":"zai","zaiToolStream":true},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":"high","medium":"high","high":"high","xhigh":"max"},
 		input: ["text"],
@@ -101,7 +101,7 @@ export const ZAI_CODING_CN_MODELS = {
 		api: "openai-completions",
 		provider: "zai-coding-cn",
 		baseUrl: "https://open.bigmodel.cn/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai","zaiToolStream":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"thinkingFormat":"zai","zaiToolStream":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -10,7 +10,7 @@ export const ZAI_MODELS = {
 		api: "openai-completions",
 		provider: "zai",
 		baseUrl: "https://api.z.ai/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai"},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"thinkingFormat":"zai"},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -28,7 +28,7 @@ export const ZAI_MODELS = {
 		api: "openai-completions",
 		provider: "zai",
 		baseUrl: "https://api.z.ai/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai","zaiToolStream":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"thinkingFormat":"zai","zaiToolStream":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -46,7 +46,7 @@ export const ZAI_MODELS = {
 		api: "openai-completions",
 		provider: "zai",
 		baseUrl: "https://api.z.ai/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai","zaiToolStream":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"thinkingFormat":"zai","zaiToolStream":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -64,7 +64,7 @@ export const ZAI_MODELS = {
 		api: "openai-completions",
 		provider: "zai",
 		baseUrl: "https://api.z.ai/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai","zaiToolStream":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"thinkingFormat":"zai","zaiToolStream":true},
 		reasoning: true,
 		input: ["text"],
 		cost: {
@@ -82,7 +82,7 @@ export const ZAI_MODELS = {
 		api: "openai-completions",
 		provider: "zai",
 		baseUrl: "https://api.z.ai/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai","supportsReasoningEffort":true,"zaiToolStream":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":true,"thinkingFormat":"zai","zaiToolStream":true},
 		reasoning: true,
 		thinkingLevelMap: {"minimal":null,"low":"high","medium":"high","high":"high","xhigh":"max"},
 		input: ["text"],
@@ -101,7 +101,7 @@ export const ZAI_MODELS = {
 		api: "openai-completions",
 		provider: "zai",
 		baseUrl: "https://api.z.ai/api/coding/paas/v4",
-		compat: {"supportsDeveloperRole":false,"thinkingFormat":"zai","zaiToolStream":true},
+		compat: {"supportsStore":false,"supportsDeveloperRole":false,"supportsReasoningEffort":false,"thinkingFormat":"zai","zaiToolStream":true},
 		reasoning: true,
 		input: ["text", "image"],
 		cost: {
@@ -99,6 +99,7 @@ export type Transport = "sse" | "websocket" | "websocket-cached" | "auto";

 /** Provider-scoped environment overrides. Values take precedence over process.env. */
 export type ProviderEnv = Record<string, string>;
+export type ProviderHeaders = Record<string, string | null>;

 export interface ProviderResponse {
 	status: number;
@@ -142,8 +143,9 @@ export interface StreamOptions {
 	 * On AWS Bedrock these are injected via a Smithy `build`-step middleware so
 	 * they are covered by SigV4 signing; reserved headers (`x-amz-*`,
 	 * `authorization`, `host`) are silently ignored to preserve SigV4 / bearer auth.
+	 * A null value suppresses a provider/API default header with the same name.
 	 */
-	headers?: Record<string, string>;
+	headers?: ProviderHeaders;
 	/**
 	 * HTTP request timeout in milliseconds for providers/SDKs that support it.
 	 * For example, OpenAI and Anthropic SDK clients default to 10 minutes.
@@ -256,8 +258,9 @@ export interface ImagesOptions {
 	/**
 	 * Optional custom HTTP headers to include in API requests.
 	 * Merged with provider defaults; can override default headers.
+	 * A null value suppresses a provider/API default header with the same name.
 	 */
-	headers?: Record<string, string>;
+	headers?: ProviderHeaders;
 	/**
 	 * HTTP request timeout in milliseconds for providers/SDKs that support it.
 	 */
@@ -1,3 +1,5 @@
+import type { ProviderHeaders } from "../types.ts";
+
 export function headersToRecord(headers: Headers): Record<string, string> {
 	const result: Record<string, string> = {};
 	for (const [key, value] of headers.entries()) {
@@ -5,3 +7,12 @@ export function headersToRecord(headers: Headers): Record<string, string> {
 	}
 	return result;
 }
+
+export function providerHeadersToRecord(headers: ProviderHeaders | undefined): Record<string, string> | undefined {
+	if (!headers) return undefined;
+	const result: Record<string, string> = {};
+	for (const [key, value] of Object.entries(headers)) {
+		if (value !== null) result[key] = value;
+	}
+	return Object.keys(result).length > 0 ? result : undefined;
+}