pi-free 2.0.12 → 2.0.14

package/README.md

@@ -20,7 +20,7 @@ When you install pi-free, it:
 
 3. **Filters to show only free models by default** for providers that expose pricing — You see only the models that cost $0 to use. Paid models are hidden until you explicitly toggle them on.
 
-4. **Provides per-provider toggle commands** — Run `/toggle-{provider}` (e.g., `/toggle-kilo`, `/toggle-opencode`) to switch between free-only mode and showing all models including paid ones. Changes apply immediately and your preference is saved for the next Pi restart.
+4. **Provides per-provider toggle commands** — Run `/toggle-{provider}` (e.g., `/toggle-kilo`) to switch between free-only mode and showing all models including paid ones. Changes apply immediately and your preference is saved for the next Pi restart.
 
 5. **Handles authentication for you** — OAuth flows (Kilo, Cline) open your browser automatically; API keys are read from `~/.pi/free.json` or environment variables
 
@@ -46,7 +46,6 @@ Free models are shown by default — look for the provider prefixes:
 
 **✅ Free Models (no payment required):**
 
-- `opencode/` — OpenCode models (no setup required; toggle with `/toggle-opencode`)
 - `kilo/` — Kilo models (free models available immediately, more after `/login kilo`)
 - `openrouter/` — OpenRouter models (free account required)
 - `cline/` — Cline models (run `/login cline` to use)
@@ -65,6 +64,7 @@ Free models are shown by default — look for the provider prefixes:
 - `codestral/` — Codestral via Mistral (free Experiment plan: 2 req/min, 1B tokens/month)
 - `deepinfra/` — DeepInfra inference cloud ($5 one-time trial credit, no credit card)
 - `novita/` — Novita AI (100+ open-source models, OpenAI-compatible, 3 free models)
+- `routeway/` — Routeway AI gateway (OpenAI-compatible, `:free` models)
 
 > **Note:** Paid providers may occasionally offer free models or promotional credits. The `isFreeModel` helper automatically detects free models based on provider pricing data or model names containing "free". For providers that don't expose pricing (like CrofAI), only models with "free" in their names are marked as free.
 
@@ -75,7 +75,6 @@ Free models are shown by default — look for the provider prefixes:
 - `cerebras/` — Cerebras models (when `CEREBRAS_API_KEY` set)
 - `xai/` — xAI models (when `XAI_API_KEY` set)
 - `huggingface/` — Hugging Face models (when `HF_TOKEN` set)
-- `opencode/` — OpenCode models (fetched from opencode.ai/zen/v1, when `OPENCODE_API_KEY` set)
 - `openrouter/` — OpenRouter models (fetched from openrouter.ai, when `OPENROUTER_API_KEY` set)
 - `fastrouter/` — FastRouter models (always discovered, 170+ models, no auth for listing)
 
@@ -86,7 +85,6 @@ Free models are shown by default — look for the provider prefixes:
 Want to see paid models too? Run the toggle command for your provider:
 
 ```
-/toggle-opencode   # Toggle OpenCode (✅ offers free models)
 /toggle-kilo       # Toggle Kilo (✅ offers free models)
 /toggle-openrouter # Toggle OpenRouter (✅ offers free models)
 /toggle-cline      # Toggle Cline (✅ offers free models)
@@ -105,6 +103,7 @@ Want to see paid models too? Run the toggle command for your provider:
 /toggle-sambanova # Toggle SambaNova (🔄 freemium)
 /toggle-llm7      # Toggle LLM7 (✅ free gateway)
 /toggle-novita    # Toggle Novita AI (💳 paid — 3 free models)
+/toggle-routeway  # Toggle Routeway AI (💳 paid — has :free models)
 /toggle-fastrouter # Toggle FastRouter (🔧 dynamic — always discovered)
 ```
 
@@ -114,10 +113,6 @@ Want to see paid models too? Run the toggle command for your provider:
 - **🔧 Dynamic providers** show all fetched models by default — the toggle filters the list when you have an API key configured
 - **Freemium providers** show all models by default; you manage your usage limits via their dashboards
 
-You'll see a notification like: `opencode: showing free models` or `opencode: showing all models`
-
-**Note:** Built-in provider toggles such as OpenCode and OpenRouter update in the current session — no restart needed.
-
 ### 4. Add API keys for more providers (optional)
 
 Some providers require a free account or API key.
@@ -139,7 +134,8 @@ Add your API keys to this file:
   "sambanova_api_key": "...",
   "llm7_api_key": "...",
   "zenmux_api_key": "...",
-  "crofai_api_key": "..."
+  "crofai_api_key": "...",
+  "routeway_api_key": "sk-..."
 }
 ```
 
@@ -204,7 +200,7 @@ Providers have different pricing models. pi-free handles them all:
 
 **Provider types:**
 
-- ✅ **Free providers** (OpenCode, Kilo, Cline) — Toggle between free-only vs paid models
+- ✅ **Free providers** (Kilo, Cline) — Toggle between free-only vs paid models
 - 🔄 **Freemium** (NVIDIA, Ollama) — Free tier with limits, toggle shows all
 - 🔧 **Dynamic API** (Mistral, Groq, Cerebras, xAI) — Fetched when API key configured, toggle filters the list
 
@@ -219,17 +215,6 @@ Authentication is handled automatically:
 
 ## Using Free Models (No Setup Required)
 
-### OpenCode
-
-Works immediately with zero setup:
-
-1. Press `Ctrl+L`
-2. Search for `opencode/`
-3. Pick any model (e.g., `opencode/big-pickle`)
-4. Start chatting
-
-No account, no API key, no OAuth. Run `/toggle-opencode` to switch between free and paid OpenCode models.
-
 ### Kilo (free models, more after login)
 
 Kilo shows free models immediately. To unlock all models, authenticate with Kilo's free OAuth:
@@ -450,7 +435,6 @@ Each provider has toggle commands to switch between free and all models:
 
 | Command                 | Action                                                   |
 | ----------------------- | -------------------------------------------------------- |
-| `/toggle-opencode`      | Toggle between free/all OpenCode models                  |
 | `/toggle-kilo`          | Toggle between free/all Kilo models                      |
 | `/toggle-openrouter`    | Toggle between free/all OpenRouter models                |
 | `/toggle-cline`         | Toggle between free/all Cline models                     |
@@ -477,7 +461,7 @@ Each provider has toggle commands to switch between free and all models:
 - **For 🔄 freemium providers**: Shows all models by default; toggle switches between filtered and full list
 - **For 🔧 dynamic API providers**: Filters the model list when you have an API key configured
 - **Persists your preference** to `~/.pi/free.json` for next startup
-- Shows a notification: "opencode: showing free models" or "opencode: showing all models"
+
 
 ### Probe Commands (Health Check)
 

package/config.ts

@@ -17,6 +17,7 @@ export {
 	PROVIDER_MODAL,
 	PROVIDER_NVIDIA,
 	PROVIDER_QWEN,
+	PROVIDER_ROUTEWAY,
 } from "./constants.ts";
 import { createLogger } from "./lib/logger.ts";
 
@@ -33,6 +34,7 @@ interface PiFreeConfig {
 	sambanova_api_key?: string;
 	together_api_key?: string;
 	novita_api_key?: string;
+	routeway_api_key?: string;
 	fastrouter_api_key?: string;
 	kilo_free_only?: boolean;
 	hidden_models?: string[];
@@ -48,6 +50,7 @@ interface PiFreeConfig {
 	sambanova_show_paid?: boolean;
 	together_show_paid?: boolean;
 	novita_show_paid?: boolean;
+	routeway_show_paid?: boolean;
 	fastrouter_show_paid?: boolean;
 	openrouter_show_paid?: boolean;
 	opencode_show_paid?: boolean;
@@ -64,6 +67,7 @@ const CONFIG_TEMPLATE: PiFreeConfig = {
 	sambanova_api_key: "",
 	together_api_key: "",
 	novita_api_key: "",
+	routeway_api_key: "",
 	fastrouter_api_key: "",
 
 	kilo_free_only: false,
@@ -80,6 +84,7 @@ const CONFIG_TEMPLATE: PiFreeConfig = {
 	sambanova_show_paid: false,
 	together_show_paid: false,
 	novita_show_paid: false,
+	routeway_show_paid: false,
 	fastrouter_show_paid: false,
 	openrouter_show_paid: false,
 	opencode_show_paid: false,
@@ -222,6 +227,10 @@ export function getNovitaShowPaid(): boolean {
 	return resolveBool("NOVITA_SHOW_PAID", loadConfigFile().novita_show_paid);
 }
 
+export function getRoutewayShowPaid(): boolean {
+	return resolveBool("ROUTEWAY_SHOW_PAID", loadConfigFile().routeway_show_paid);
+}
+
 export function getFastrouterShowPaid(): boolean {
 	return resolveBool(
 		"FASTROUTER_SHOW_PAID",
@@ -266,6 +275,8 @@ export function getProviderShowPaid(providerId: string): boolean {
 			return getTogetherShowPaid();
 		case "novita":
 			return getNovitaShowPaid();
+		case "routeway":
+			return getRoutewayShowPaid();
 		case "fastrouter":
 			return getFastrouterShowPaid();
 		case "ollama-cloud":
@@ -331,6 +342,10 @@ export function getNovitaApiKey(): string | undefined {
 	return resolve("NOVITA_API_KEY", loadConfigFile().novita_api_key);
 }
 
+export function getRoutewayApiKey(): string | undefined {
+	return resolve("ROUTEWAY_API_KEY", loadConfigFile().routeway_api_key);
+}
+
 export function getFastrouterApiKey(): string | undefined {
 	return resolve("FASTROUTER_API_KEY", loadConfigFile().fastrouter_api_key);
 }

package/constants.ts

@@ -23,6 +23,7 @@ export const PROVIDER_DEEPINFRA = "deepinfra";
 export const PROVIDER_SAMBANOVA = "sambanova";
 export const PROVIDER_TOGETHER = "together";
 export const PROVIDER_NOVITA = "novita";
+export const PROVIDER_ROUTEWAY = "routeway";
 
 export const ALL_UNIQUE_PROVIDERS = [
 	PROVIDER_KILO,
@@ -40,6 +41,7 @@ export const ALL_UNIQUE_PROVIDERS = [
 	PROVIDER_SAMBANOVA,
 	PROVIDER_TOGETHER,
 	PROVIDER_NOVITA,
+	PROVIDER_ROUTEWAY,
 ] as const;
 
 // =============================================================================
@@ -62,6 +64,7 @@ export const BASE_URL_DEEPINFRA = "https://api.deepinfra.com/v1/openai";
 export const BASE_URL_SAMBANOVA = "https://api.sambanova.ai/v1";
 export const BASE_URL_TOGETHER = "https://api.together.xyz/v1";
 export const BASE_URL_NOVITA = "https://api.novita.ai/openai/v1";
+export const BASE_URL_ROUTEWAY = "https://api.routeway.ai/v1";
 
 /** Cline fetches free models from OpenRouter */
 export const BASE_URL_OPENROUTER = "https://openrouter.ai/api/v1";

package/index.ts

@@ -14,6 +14,7 @@
  * - DeepInfra: AI inference cloud ($5 trial credit)
  * - SambaNova: Fast inference on RDU hardware (free tier, no credit card)
  * - Together: Fast inference on 200+ open-source models ($1 trial credit)
+ * - Routeway: OpenAI-compatible gateway with free `:free` models
  * - LLM7: AI gateway (free default/fast selectors)
  */
 
@@ -24,6 +25,13 @@ import {
 	processQuotaResponse,
 	formatQuotaStatus,
 } from "./lib/quota-monitor.ts";
+import {
+	startModelCall,
+	recordModelCall,
+	getAllTelemetry,
+	getTelemetryPath,
+	clearTelemetry,
+} from "./lib/telemetry.ts";
 import {
 	applyGlobalFilter,
 	getGlobalFreeOnly,
@@ -41,6 +49,7 @@ import deepinfra from "./providers/deepinfra/deepinfra.ts";
 import sambanova from "./providers/sambanova/sambanova.ts";
 import together from "./providers/together/together.ts";
 import novita from "./providers/novita/novita.ts";
+import routeway from "./providers/routeway/routeway.ts";
 import nvidia from "./providers/nvidia/nvidia.ts";
 import ollama from "./providers/ollama/ollama.ts";
 import zenmux from "./providers/zenmux/zenmux.ts";
@@ -145,6 +154,64 @@ function setupGlobalCommands(pi: ExtensionAPI) {
 			ctx.ui.notify(lines.join("\n"), "info");
 		},
 	});
+
+	// /telemetry — Show model telemetry data
+	pi.registerCommand("free-telemetry", {
+		description:
+			"Show real-world performance data for free models (tokens/s, latency, success rate)",
+		handler: async (_args, ctx) => {
+			const allTelemetry = getAllTelemetry();
+			const entries = Object.entries(allTelemetry);
+
+			if (entries.length === 0) {
+				ctx.ui.notify(
+					"No telemetry data yet. Use some free models first!",
+					"info",
+				);
+				return;
+			}
+
+			// Sort by total calls descending
+			entries.sort((a, b) => b[1].totalCalls - a[1].totalCalls);
+
+			const lines = ["📊 Model Telemetry:", ""];
+			lines.push(
+				`${`Model`.padEnd(40)} ${`Calls`.padEnd(6)} ${`OK%`.padEnd(6)} ${`Lat`.padEnd(7)} ${`tok/s`.padEnd(7)} ${`Cost`}`,
+			);
+			lines.push(`─`.repeat(75));
+
+			for (const [key, t] of entries.slice(0, 20)) {
+				const name = key.length > 38 ? key.slice(0, 35) + "..." : key;
+				const calls = String(t.totalCalls).padStart(5);
+				const ok = `${t.successRate}%`.padStart(5);
+				const lat =
+					t.avgLatencyMs > 0
+						? `${t.avgLatencyMs}ms`.padStart(6)
+						: "—".padStart(6);
+				const tps =
+					t.avgTokensPerSecond > 0
+						? `${t.avgTokensPerSecond}`.padStart(6)
+						: "—".padStart(6);
+				const cost =
+					t.totalCost > 0
+						? `$${t.totalCost.toFixed(4)}`.padStart(8)
+						: "free".padStart(8);
+				lines.push(`${name.padEnd(40)} ${calls} ${ok} ${lat} ${tps} ${cost}`);
+			}
+
+			lines.push("", `File: ${getTelemetryPath()}`);
+			ctx.ui.notify(lines.join("\n"), "info");
+		},
+	});
+
+	// /clear-free-telemetry — Clear all telemetry data
+	pi.registerCommand("clear-free-telemetry", {
+		description: "Clear all model telemetry data",
+		handler: async (_args, ctx) => {
+			clearTelemetry();
+			ctx.ui.notify("Telemetry data cleared", "info");
+		},
+	});
 }
 
 // =============================================================================
@@ -183,6 +250,70 @@ function setupQuotaMonitoring(pi: ExtensionAPI) {
 	});
 }
 
+// =============================================================================
+// Model Telemetry
+// =============================================================================
+
+function setupTelemetry(pi: ExtensionAPI) {
+	// Only track telemetry for FREE models (uses same isFreeModel logic as model filtering)
+	// eslint-disable-next-line @typescript-eslint/no-explicit-any
+	(pi as any).on("before_agent_start", (_event: any, ctx: any) => {
+		if (!ctx.model) return;
+		if (!isFreeModel(ctx.model as any)) return;
+		const provider = ctx.model?.provider;
+		const model = ctx.model?.id;
+		if (provider && model) {
+			startModelCall(provider, model);
+		}
+	});
+
+	// Record telemetry when a turn completes
+	pi.on("turn_end", (event, ctx) => {
+		if (!ctx.model) return;
+		if (!isFreeModel(ctx.model as any)) return;
+
+		const msg = (
+			event as {
+				message?: {
+					role?: string;
+					model?: string;
+					usage?: {
+						input?: number;
+						output?: number;
+						totalTokens?: number;
+						cost?: { total?: number };
+					};
+					stopReason?: string;
+					errorMessage?: string;
+				};
+			}
+		).message;
+
+		if (msg?.role !== "assistant") return;
+
+		const provider = ctx.model?.provider;
+		const model = msg.model || ctx.model?.id;
+		if (!provider || !model) return;
+
+		const usage = msg.usage;
+		const inputTokens = usage?.input ?? 0;
+		const outputTokens = usage?.output ?? 0;
+		const totalTokens = usage?.totalTokens ?? inputTokens + outputTokens;
+		const cost = usage?.cost?.total ?? 0;
+		const isError = msg.stopReason === "error" || !!msg.errorMessage;
+
+		recordModelCall(
+			provider,
+			model,
+			{ input: inputTokens, output: outputTokens, totalTokens },
+			cost,
+			!isError,
+			msg.stopReason,
+			msg.errorMessage,
+		);
+	});
+}
+
 // =============================================================================
 // Main Entry Point
 // =============================================================================
@@ -197,6 +328,9 @@ export default async function piFreeEntry(pi: ExtensionAPI) {
 	// Setup quota monitoring (passive, no extra API calls)
 	setupQuotaMonitoring(pi);
 
+	// Setup model telemetry (tracks real-world performance)
+	setupTelemetry(pi);
+
 	// Load all unique providers
 	// Each provider will register itself with the global toggle system
 	await Promise.allSettled([
@@ -212,6 +346,7 @@ export default async function piFreeEntry(pi: ExtensionAPI) {
 		sambanova(pi),
 		together(pi),
 		novita(pi),
+		routeway(pi),
 	]);
 
 	// Setup dynamic built-in providers (Mistral, Groq, Cerebras, xAI, Hugging Face,

package/lib/built-in-toggle.ts

@@ -24,9 +24,18 @@ import {
 	registerWithGlobalToggle,
 } from "./registry.ts";
 import { createToggleState } from "./toggle-state.ts";
+import {
+	OPENCODE_DYNAMIC_API,
+	createOpenCodeSessionTracker,
+	createOpenCodeStreamSimple,
+	isOpenCodeProvider,
+} from "../providers/opencode-session.ts";
 
 const _logger = createLogger("built-in-toggle");
 
+// OpenCode requires per-request ids; see createOpenCodeStreamSimple().
+const _opencodeSession = createOpenCodeSessionTracker();
+
 // =============================================================================
 // Configuration
 // =============================================================================
@@ -38,6 +47,7 @@ interface BuiltInToggleConfig {
 
 const BUILT_IN_TOGGLE_PROVIDERS: BuiltInToggleConfig[] = [
 	{ id: "opencode", getShowPaid: getOpencodeShowPaid },
+	{ id: "opencode-go", getShowPaid: getOpencodeShowPaid },
 	{ id: "openrouter", getShowPaid: getOpenrouterShowPaid },
 ];
 
@@ -113,7 +123,9 @@ function tryCaptureProvider(
 	);
 	if (providerModels.length === 0) return undefined;
 
-	const allModels = providerModels.map(modelToProviderConfig);
+	const allModels = providerModels.map((m: Model<Api>) =>
+		modelToProviderConfig(m, config.id),
+	);
 	const freeModels = allModels.filter((m: ProviderModelConfig) =>
 		isFreeModel({ ...m, provider: config.id }, allModels),
 	);
@@ -126,7 +138,10 @@ function tryCaptureProvider(
 		pi.registerProvider(config.id, {
 			baseUrl,
 			apiKey: apiKeyEnv,
-			api,
+			api: isOpenCodeProvider(config.id) ? OPENCODE_DYNAMIC_API : api,
+			...(isOpenCodeProvider(config.id)
+				? { streamSimple: createOpenCodeStreamSimple(_opencodeSession) }
+				: {}),
 			models,
 		});
 	};
@@ -196,8 +211,11 @@ function registerToggleCommand(
 // Helpers
 // =============================================================================
 
-function modelToProviderConfig(m: Model<Api>): ProviderModelConfig {
-	return {
+function modelToProviderConfig(
+	m: Model<Api>,
+	providerId?: string,
+): ProviderModelConfig {
+	const base: ProviderModelConfig = {
 		id: m.id,
 		name: m.name,
 		api: m.api,
@@ -209,6 +227,14 @@ function modelToProviderConfig(m: Model<Api>): ProviderModelConfig {
 		headers: m.headers,
 		compat: (m as any).compat,
 	};
+
+	// Use a custom OpenCode API wrapper so per-request headers are regenerated
+	// for every LLM call instead of being frozen at registration time.
+	if (providerId && isOpenCodeProvider(providerId)) {
+		base.api = OPENCODE_DYNAMIC_API;
+	}
+
+	return base;
 }
 
 // =============================================================================
@@ -252,8 +278,9 @@ function setupStatusBar(
 
 function getApiKeyEnvForProvider(providerId: string): string {
 	const envMap: Record<string, string> = {
-		opencode: "OPENCODE_API_KEY",
-		openrouter: "OPENROUTER_API_KEY",
+		opencode: "$OPENCODE_API_KEY",
+		"opencode-go": "$OPENCODE_API_KEY",
+		openrouter: "$OPENROUTER_API_KEY",
 	};
-	return envMap[providerId] || `${providerId.toUpperCase()}_API_KEY`;
+	return envMap[providerId] || `$${providerId.toUpperCase()}_API_KEY`;
 }

package/lib/probe-cache.ts

@@ -0,0 +1,86 @@
+/**
+ * Provider model probe cache.
+ *
+ * Stores the last successful accessibility probe per provider/model so
+ * background cleanup can avoid spending quota on the same checks every session.
+ */
+
+import { homedir } from "node:os";
+import { join } from "node:path";
+import { createJSONStore } from "./json-persistence.ts";
+import { createLogger } from "./logger.ts";
+
+const _logger = createLogger("probe-cache");
+
+export const DEFAULT_PROBE_TTL_MS = 24 * 60 * 60 * 1000;
+
+export type ProbeStatus = "ok" | "broken";
+
+export interface ModelProbeResult {
+	modelId: string;
+	status: ProbeStatus;
+}
+
+interface ModelProbeEntry {
+	lastProbedAt: string;
+	status: ProbeStatus;
+}
+
+interface ProviderProbeCache {
+	provider: string;
+	models: Record<string, ModelProbeEntry>;
+}
+
+interface ProbeCacheData {
+	providers: Record<string, ProviderProbeCache>;
+}
+
+const CACHE_FILE = join(homedir(), ".pi", "probe-cache.json");
+const _cache = createJSONStore<ProbeCacheData>(CACHE_FILE, { providers: {} });
+
+export function getModelsDueForProbe(
+	providerId: string,
+	modelIds: string[],
+	ttlMs = DEFAULT_PROBE_TTL_MS,
+): string[] {
+	const provider = _cache.load().providers[providerId];
+	const now = Date.now();
+
+	return modelIds.filter((modelId) => {
+		const entry = provider?.models[modelId];
+		if (!entry) return true;
+
+		// Broken models are normally hidden immediately. If a user later unhides one,
+		// re-check it instead of letting a stale broken cache suppress cleanup.
+		if (entry.status === "broken") return true;
+
+		const lastProbedAt = Date.parse(entry.lastProbedAt);
+		if (!Number.isFinite(lastProbedAt)) return true;
+
+		return now - lastProbedAt >= ttlMs;
+	});
+}
+
+export function recordModelProbeResults(
+	providerId: string,
+	results: ModelProbeResult[],
+): void {
+	if (results.length === 0) return;
+
+	const data = _cache.load();
+	const provider = (data.providers[providerId] ??= {
+		provider: providerId,
+		models: {},
+	});
+	const lastProbedAt = new Date().toISOString();
+
+	for (const result of results) {
+		provider.models[result.modelId] = {
+			lastProbedAt,
+			status: result.status,
+		};
+	}
+
+	_cache.save(data);
+	_logger.debug(`Recorded ${results.length} probe results for ${providerId}`);
+}

package/lib/registry.ts

@@ -82,7 +82,12 @@ function detectPricingExposed(allModels: ProviderModelConfig[]): boolean {
  * @returns true if the model is definitively free per the provider's API
  */
 export function isFreeModel(
-	model: ProviderModelConfig & { provider?: string; _pricingKnown?: boolean },
+	model: ProviderModelConfig & {
+		provider?: string;
+		_pricingKnown?: boolean;
+		_freeKnown?: boolean;
+		_isFree?: boolean;
+	},
 	allModels?: ProviderModelConfig[],
 ): boolean {
 	return isFreeModelInternal(model, allModels);
@@ -90,9 +95,21 @@ export function isFreeModel(
 
 // Internal implementation to work around TypeScript filter callback issues
 function isFreeModelInternal(
-	model: ProviderModelConfig & { provider?: string; _pricingKnown?: boolean },
+	model: ProviderModelConfig & {
+		provider?: string;
+		_pricingKnown?: boolean;
+		_freeKnown?: boolean;
+		_isFree?: boolean;
+	},
 	allModels: ProviderModelConfig[] | undefined,
 ): boolean {
+	// Some gateways expose an authoritative free/paid flag. Prefer it over
+	// pricing because a few non-chat or preview models can report zero token
+	// prices while still not being offered as free chat models.
+	if (model._freeKnown === true) {
+		return model._isFree === true;
+	}
+
 	// Determine if pricing is exposed
 	let pricingExposed: boolean;
 
@@ -213,7 +230,12 @@ export function applyGlobalFilter(
 
 	for (const [providerId, entry] of providerRegistry) {
 		try {
-			applyFilterToProvider(providerId, entry, freeOnly, options.force === true);
+			applyFilterToProvider(
+				providerId,
+				entry,
+				freeOnly,
+				options.force === true,
+			);
 		} catch (err) {
 			_logger.error(
 				`[pi-free] Failed to apply filter to ${providerId}`,