@raingor/pi-web-switch 0.4.3 → 0.4.4

package/src/data/builtin-providers.test.ts

@@ -0,0 +1,109 @@
+import { existsSync, readFileSync, readdirSync } from "node:fs";
+import { dirname, join } from "node:path";
+import { describe, expect, it } from "vitest";
+import { BUILTIN_PROVIDERS } from "./builtin-providers";
+
+// This file is only the fallback for /api/pi/builtin-providers, which normally
+// serves pi's real catalog from @earendil-works/pi-ai. Drift between the two is
+// invisible at runtime (the fallback only kicks in when pi is missing), so these
+// tests pin the fallback to the catalog shipped with the pinned dependency.
+
+type CatalogModel = {
+  id: string;
+  name?: string;
+  reasoning?: boolean;
+  input?: string[];
+  contextWindow?: number;
+  maxTokens?: number;
+  cost?: { input?: number; output?: number; cacheRead?: number; cacheWrite?: number };
+};
+
+function catalogDir(): string | null {
+  let dir = dirname(new URL(import.meta.url).pathname);
+  for (let i = 0; i < 6; i++) {
+    const candidate = join(dir, "node_modules", "@earendil-works", "pi-ai", "dist", "providers");
+    if (existsSync(join(candidate, "data"))) return candidate;
+    dir = dirname(dir);
+  }
+  return null;
+}
+
+const dir = catalogDir();
+
+/** Flatten one provider's data file into an id -> model map, as pi-reader does. */
+function readProvider(providerId: string): Record<string, CatalogModel> | null {
+  if (!dir) return null;
+  const file = join(dir, "data", `${providerId}.json`);
+  if (!existsSync(file)) return null;
+  const data = JSON.parse(readFileSync(file, "utf-8")) as Record<string, Record<string, CatalogModel>>;
+  const out: Record<string, CatalogModel> = {};
+  for (const api of Object.keys(data)) {
+    for (const model of Object.values(data[api] ?? {})) {
+      if (model?.id) out[model.id] = model;
+    }
+  }
+  return out;
+}
+
+describe.skipIf(!dir)("builtin providers match the pi-ai catalog", () => {
+  it("every provider id exists in the catalog", () => {
+    const available = new Set(
+      readdirSync(join(dir!, "data"))
+        .filter((f) => f.endsWith(".json") && !f.startsWith("."))
+        .map((f) => f.replace(/\.json$/, ""))
+    );
+    const unknown = BUILTIN_PROVIDERS.map((p) => p.id).filter((id) => !available.has(id));
+    expect(unknown).toEqual([]);
+  });
+
+  it.each(BUILTIN_PROVIDERS.map((p) => [p.id, p] as const))(
+    "%s models match the catalog",
+    (providerId, provider) => {
+      const catalog = readProvider(providerId);
+      expect(catalog, `${providerId}.json missing`).toBeTruthy();
+
+      for (const model of provider.models) {
+        const ref = catalog![model.id] as CatalogModel | undefined;
+        expect(ref, `${providerId}/${model.id} is not in the catalog`).toBeDefined();
+        if (!ref) continue;
+        expect(model.contextWindow, `${providerId}/${model.id} contextWindow`).toBe(ref.contextWindow);
+        expect(model.maxTokens, `${providerId}/${model.id} maxTokens`).toBe(ref.maxTokens);
+        expect(model.reasoning, `${providerId}/${model.id} reasoning`).toBe(!!ref.reasoning);
+        expect(model.input, `${providerId}/${model.id} input`).toEqual(ref.input ?? ["text"]);
+        // Entry pricing tier only — cost.tiers surcharges have no Model field.
+        expect(model.cost?.input, `${providerId}/${model.id} cost.input`).toBe(ref.cost?.input ?? 0);
+        expect(model.cost?.output, `${providerId}/${model.id} cost.output`).toBe(ref.cost?.output ?? 0);
+        expect(model.cost?.cacheRead, `${providerId}/${model.id} cost.cacheRead`).toBe(ref.cost?.cacheRead ?? 0);
+        expect(model.cost?.cacheWrite, `${providerId}/${model.id} cost.cacheWrite`).toBe(ref.cost?.cacheWrite ?? 0);
+      }
+    }
+  );
+});
+
+describe("builtin providers shape", () => {
+  it("provider ids are unique", () => {
+    const ids = BUILTIN_PROVIDERS.map((p) => p.id);
+    expect(new Set(ids).size).toBe(ids.length);
+  });
+
+  it("model ids are unique within each provider", () => {
+    for (const provider of BUILTIN_PROVIDERS) {
+      const ids = provider.models.map((m) => m.id);
+      expect(new Set(ids).size, provider.id).toBe(ids.length);
+    }
+  });
+
+  it("maxTokens never exceeds contextWindow", () => {
+    for (const provider of BUILTIN_PROVIDERS) {
+      for (const model of provider.models) {
+        expect(model.maxTokens!, `${provider.id}/${model.id}`).toBeLessThanOrEqual(model.contextWindow!);
+      }
+    }
+  });
+
+  it("hasAuth agrees with authMethod", () => {
+    for (const provider of BUILTIN_PROVIDERS) {
+      expect(provider.hasAuth, provider.id).toBe(provider.authMethod !== "none");
+    }
+  });
+});

package/src/data/builtin-providers.ts

@@ -1,6 +1,13 @@
-// ─── Built-in Providers ───────────────────────────────────
-// These are hardcoded in pi's Rust source and not exposed as config files.
-// We maintain this list to display known providers in the UI.
+// ─── Built-in Providers (static fallback) ─────────────────
+// pi's real builtin catalog lives in @earendil-works/pi-ai as
+// dist/providers/data/*.json and is served at /api/pi/builtin-providers via
+// readBuiltinCatalog(). This list is only the fallback used when pi cannot be
+// located on the machine, so it must stay consistent with that catalog.
+//
+// Generated from the pi-ai catalog shipped with this repo's pinned
+// @earendil-works/pi-ai. Costs are USD per 1M tokens; only the entry pricing
+// tier is kept (pi-ai's `cost.tiers` long-context surcharges are dropped
+// because Model has no field for them).
 
 import type { Provider } from "@/types";
 
@@ -10,29 +17,31 @@ export const BUILTIN_PROVIDERS: Provider[] = [
     name: "Anthropic",
     type: "builtin",
     api: "anthropic-messages",
-    baseUrl: "https://api.anthropic.com/v1",
+    baseUrl: "https://api.anthropic.com",
     hasAuth: true,
     authMethod: "env",
     models: [
-      { id: "claude-sonnet-4", name: "Claude 4 Sonnet", reasoning: true, input: ["text", "image"], contextWindow: 200000, maxTokens: 16384, cost: { input: 3, output: 15, cacheRead: 0.3, cacheWrite: 3.75 }, enabled: true },
-      { id: "claude-sonnet-4-5", name: "Claude 4.5 Sonnet", reasoning: true, input: ["text", "image"], contextWindow: 200000, maxTokens: 16384, cost: { input: 3, output: 15, cacheRead: 0.3, cacheWrite: 3.75 }, enabled: false },
-      { id: "claude-opus-4", name: "Claude 4 Opus", reasoning: true, input: ["text", "image"], contextWindow: 200000, maxTokens: 32000, cost: { input: 15, output: 75, cacheRead: 1.5, cacheWrite: 18.75 }, enabled: false },
-      { id: "claude-haiku-3-5", name: "Claude 3.5 Haiku", reasoning: false, input: ["text", "image"], contextWindow: 200000, maxTokens: 8192, cost: { input: 0.8, output: 4, cacheRead: 0.08, cacheWrite: 1 }, enabled: true },
+      { id: "claude-opus-5", name: "Claude Opus 5", reasoning: true, input: ["text", "image"], contextWindow: 1000000, maxTokens: 128000, cost: { input: 5, output: 25, cacheRead: 0.5, cacheWrite: 6.25 }, enabled: false },
+      { id: "claude-sonnet-5", name: "Claude Sonnet 5", reasoning: true, input: ["text", "image"], contextWindow: 1000000, maxTokens: 128000, cost: { input: 2, output: 10, cacheRead: 0.2, cacheWrite: 2.5 }, enabled: true },
+      { id: "claude-opus-4-5", name: "Claude Opus 4.5 (latest)", reasoning: true, input: ["text", "image"], contextWindow: 200000, maxTokens: 64000, cost: { input: 5, output: 25, cacheRead: 0.5, cacheWrite: 6.25 }, enabled: true },
+      { id: "claude-sonnet-4-5", name: "Claude Sonnet 4.5 (latest)", reasoning: true, input: ["text", "image"], contextWindow: 1000000, maxTokens: 64000, cost: { input: 3, output: 15, cacheRead: 0.3, cacheWrite: 3.75 }, enabled: true },
+      { id: "claude-haiku-4-5", name: "Claude Haiku 4.5 (latest)", reasoning: true, input: ["text", "image"], contextWindow: 200000, maxTokens: 64000, cost: { input: 1, output: 5, cacheRead: 0.1, cacheWrite: 1.25 }, enabled: true },
     ],
   },
   {
     id: "openai",
     name: "OpenAI",
     type: "builtin",
-    api: "openai-completions",
+    api: "openai-responses",
     baseUrl: "https://api.openai.com/v1",
     hasAuth: true,
     authMethod: "env",
     models: [
-      { id: "gpt-4o", name: "GPT-4o", reasoning: false, input: ["text", "image"], contextWindow: 128000, maxTokens: 16384, cost: { input: 2.5, output: 10, cacheRead: 1.25, cacheWrite: 3.75 }, enabled: true },
-      { id: "gpt-4o-mini", name: "GPT-4o Mini", reasoning: false, input: ["text", "image"], contextWindow: 128000, maxTokens: 16384, cost: { input: 0.15, output: 0.6, cacheRead: 0.075, cacheWrite: 0.225 }, enabled: true },
-      { id: "gpt-5.1", name: "GPT-5.1", reasoning: true, input: ["text", "image"], contextWindow: 256000, maxTokens: 65536, cost: { input: 10, output: 40, cacheRead: 5, cacheWrite: 10 }, enabled: false },
-      { id: "o3-mini", name: "o3-mini", reasoning: true, input: ["text"], contextWindow: 200000, maxTokens: 100000, cost: { input: 1.1, output: 4.4, cacheRead: 0.55, cacheWrite: 1.65 }, enabled: false },
+      { id: "gpt-5.6-sol", name: "GPT-5.6 Sol", reasoning: true, input: ["text", "image"], contextWindow: 272000, maxTokens: 128000, cost: { input: 5, output: 30, cacheRead: 0.5, cacheWrite: 6.25 }, enabled: true },
+      { id: "gpt-5.6-terra", name: "GPT-5.6 Terra", reasoning: true, input: ["text", "image"], contextWindow: 272000, maxTokens: 128000, cost: { input: 2.5, output: 15, cacheRead: 0.25, cacheWrite: 3.125 }, enabled: true },
+      { id: "gpt-5.6-luna", name: "GPT-5.6 Luna", reasoning: true, input: ["text", "image"], contextWindow: 272000, maxTokens: 128000, cost: { input: 1, output: 6, cacheRead: 0.1, cacheWrite: 1.25 }, enabled: false },
+      { id: "gpt-5.1", name: "GPT-5.1", reasoning: true, input: ["text", "image"], contextWindow: 400000, maxTokens: 128000, cost: { input: 1.25, output: 10, cacheRead: 0.125, cacheWrite: 0 }, enabled: true },
+      { id: "o3-mini", name: "o3-mini", reasoning: true, input: ["text"], contextWindow: 200000, maxTokens: 100000, cost: { input: 1.1, output: 4.4, cacheRead: 0.55, cacheWrite: 0 }, enabled: false },
     ],
   },
   {
@@ -40,52 +49,60 @@ export const BUILTIN_PROVIDERS: Provider[] = [
     name: "DeepSeek",
     type: "builtin",
     api: "openai-completions",
-    baseUrl: "https://api.deepseek.com/v1",
+    baseUrl: "https://api.deepseek.com",
     hasAuth: true,
     authMethod: "env",
     models: [
-      { id: "deepseek-chat", name: "DeepSeek V3", reasoning: false, input: ["text"], contextWindow: 128000, maxTokens: 8192, cost: { input: 0.27, output: 1.1, cacheRead: 0.07, cacheWrite: 0.27 }, enabled: true },
-      { id: "deepseek-reasoner", name: "DeepSeek R1", reasoning: true, input: ["text"], contextWindow: 128000, maxTokens: 8192, cost: { input: 0.55, output: 2.19, cacheRead: 0.14, cacheWrite: 0.55 }, enabled: true },
+      { id: "deepseek-v4-flash", name: "DeepSeek V4 Flash", reasoning: true, input: ["text"], contextWindow: 1000000, maxTokens: 384000, cost: { input: 0.14, output: 0.28, cacheRead: 0.0028, cacheWrite: 0 }, enabled: true },
+      { id: "deepseek-v4-pro", name: "DeepSeek V4 Pro", reasoning: true, input: ["text"], contextWindow: 1000000, maxTokens: 384000, cost: { input: 0.435, output: 0.87, cacheRead: 0.003625, cacheWrite: 0 }, enabled: true },
     ],
   },
   {
-    id: "opencode",
-    name: "OpenCode",
+    id: "google",
+    name: "Google",
     type: "builtin",
-    api: "openai-completions",
+    api: "google-generative-ai",
+    baseUrl: "https://generativelanguage.googleapis.com/v1beta",
     hasAuth: true,
-    authMethod: "file",
+    authMethod: "env",
     models: [
-      { id: "deepseek-v4-flash-free", name: "DeepSeek V4 Flash (Free)", reasoning: false, input: ["text"], contextWindow: 128000, maxTokens: 8192, cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 }, enabled: true },
-      { id: "deepseek-v4-flash", name: "DeepSeek V4 Flash", reasoning: false, input: ["text"], contextWindow: 128000, maxTokens: 8192, cost: { input: 0.3, output: 0.6, cacheRead: 0.15, cacheWrite: 0.3 }, enabled: true },
+      { id: "gemini-3.1-pro-preview", name: "Gemini 3.1 Pro Preview", reasoning: true, input: ["text", "image"], contextWindow: 1048576, maxTokens: 65536, cost: { input: 2, output: 12, cacheRead: 0.2, cacheWrite: 0 }, enabled: false },
+      { id: "gemini-3.5-flash", name: "Gemini 3.5 Flash", reasoning: true, input: ["text", "image"], contextWindow: 1048576, maxTokens: 65536, cost: { input: 1.5, output: 9, cacheRead: 0.15, cacheWrite: 0 }, enabled: true },
+      { id: "gemini-2.5-pro", name: "Gemini 2.5 Pro", reasoning: true, input: ["text", "image"], contextWindow: 1048576, maxTokens: 65536, cost: { input: 1.25, output: 10, cacheRead: 0.125, cacheWrite: 0 }, enabled: false },
+      { id: "gemini-2.5-flash", name: "Gemini 2.5 Flash", reasoning: true, input: ["text", "image"], contextWindow: 1048576, maxTokens: 65536, cost: { input: 0.3, output: 2.5, cacheRead: 0.03, cacheWrite: 0 }, enabled: true },
     ],
   },
   {
-    id: "opencode-go",
-    name: "OpenCode Go",
+    id: "opencode",
+    name: "OpenCode Zen",
     type: "builtin",
     api: "openai-completions",
+    baseUrl: "https://opencode.ai/zen/v1",
     hasAuth: true,
     authMethod: "file",
     models: [
-      { id: "deepseek-v4-flash", name: "DeepSeek V4 Flash", reasoning: false, input: ["text"], contextWindow: 128000, maxTokens: 8192, cost: { input: 0.3, output: 0.6, cacheRead: 0.15, cacheWrite: 0.3 }, enabled: true },
-      { id: "deepseek-v4-pro", name: "DeepSeek V4 Pro", reasoning: true, input: ["text"], contextWindow: 128000, maxTokens: 8192, cost: { input: 2, output: 8, cacheRead: 1, cacheWrite: 2 }, enabled: true },
-      { id: "glm-5.1", name: "GLM 5.1", reasoning: false, input: ["text"], contextWindow: 128000, maxTokens: 8192, cost: { input: 0.5, output: 2, cacheRead: 0.25, cacheWrite: 0.5 }, enabled: true },
-      { id: "qwen3.7-max", name: "Qwen 3.7 Max", reasoning: true, input: ["text"], contextWindow: 128000, maxTokens: 8192, cost: { input: 1.5, output: 6, cacheRead: 0.75, cacheWrite: 1.5 }, enabled: true },
-      { id: "mimo-v2.5", name: "MiMo V2.5", reasoning: true, input: ["text"], contextWindow: 128000, maxTokens: 8192, cost: { input: 1.2, output: 4.8, cacheRead: 0.6, cacheWrite: 1.2 }, enabled: true },
+      { id: "deepseek-v4-flash-free", name: "DeepSeek V4 Flash Free", reasoning: true, input: ["text"], contextWindow: 200000, maxTokens: 128000, cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 }, enabled: true },
+      { id: "deepseek-v4-flash", name: "DeepSeek V4 Flash", reasoning: true, input: ["text"], contextWindow: 1000000, maxTokens: 384000, cost: { input: 0.14, output: 0.28, cacheRead: 0.028, cacheWrite: 0 }, enabled: true },
+      { id: "claude-sonnet-5", name: "Claude Sonnet 5", reasoning: true, input: ["text", "image"], contextWindow: 1000000, maxTokens: 128000, cost: { input: 2, output: 10, cacheRead: 0.2, cacheWrite: 2.5 }, enabled: true },
+      { id: "gpt-5", name: "GPT-5", reasoning: true, input: ["text", "image"], contextWindow: 400000, maxTokens: 128000, cost: { input: 1.07, output: 8.5, cacheRead: 0.107, cacheWrite: 0 }, enabled: false },
+      { id: "glm-5.2", name: "GLM-5.2", reasoning: true, input: ["text"], contextWindow: 1000000, maxTokens: 131072, cost: { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 }, enabled: true },
     ],
   },
   {
-    id: "google",
-    name: "Google Gemini",
+    id: "opencode-go",
+    name: "OpenCode Zen Go",
     type: "builtin",
-    api: "google-generative-ai",
-    baseUrl: "https://generativelanguage.googleapis.com/v1beta",
+    api: "openai-completions",
+    baseUrl: "https://opencode.ai/zen/go/v1",
     hasAuth: true,
-    authMethod: "env",
+    authMethod: "file",
     models: [
-      { id: "gemini-2.5-flash", name: "Gemini 2.5 Flash", reasoning: false, input: ["text", "image"], contextWindow: 1048576, maxTokens: 65536, cost: { input: 0.15, output: 0.6, cacheRead: 0.075, cacheWrite: 0.15 }, enabled: true },
-      { id: "gemini-2.5-pro", name: "Gemini 2.5 Pro", reasoning: true, input: ["text", "image"], contextWindow: 1048576, maxTokens: 65536, cost: { input: 1.25, output: 10, cacheRead: 0.625, cacheWrite: 1.25 }, enabled: false },
+      { id: "deepseek-v4-flash", name: "DeepSeek V4 Flash", reasoning: true, input: ["text"], contextWindow: 1000000, maxTokens: 384000, cost: { input: 0.14, output: 0.28, cacheRead: 0.0028, cacheWrite: 0 }, enabled: true },
+      { id: "deepseek-v4-pro", name: "DeepSeek V4 Pro", reasoning: true, input: ["text"], contextWindow: 1000000, maxTokens: 384000, cost: { input: 0.435, output: 0.87, cacheRead: 0.003625, cacheWrite: 0 }, enabled: true },
+      { id: "glm-5.2", name: "GLM-5.2", reasoning: true, input: ["text"], contextWindow: 1000000, maxTokens: 131072, cost: { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 }, enabled: true },
+      { id: "qwen3.7-max", name: "Qwen3.7 Max", reasoning: true, input: ["text"], contextWindow: 1000000, maxTokens: 65536, cost: { input: 2.5, output: 7.5, cacheRead: 0.5, cacheWrite: 3.125 }, enabled: true },
+      { id: "minimax-m3", name: "MiniMax-M3", reasoning: true, input: ["text", "image"], contextWindow: 1000000, maxTokens: 131072, cost: { input: 0.3, output: 1.2, cacheRead: 0.06, cacheWrite: 0 }, enabled: true },
+      { id: "kimi-k3", name: "Kimi K3 (2x usage)", reasoning: true, input: ["text", "image"], contextWindow: 1048576, maxTokens: 131072, cost: { input: 3, output: 15, cacheRead: 0.3, cacheWrite: 0 }, enabled: true },
     ],
   },
   {
@@ -97,8 +114,9 @@ export const BUILTIN_PROVIDERS: Provider[] = [
     hasAuth: false,
     authMethod: "none",
     models: [
-      { id: "openrouter/anthropic/claude-sonnet-4", name: "Claude 4 Sonnet (OpenRouter)", reasoning: true, input: ["text", "image"], contextWindow: 200000, maxTokens: 16384, cost: { input: 3, output: 15, cacheRead: 0.3, cacheWrite: 3.75 }, enabled: false },
-      { id: "openrouter/deepseek/deepseek-r1", name: "DeepSeek R1 (OpenRouter)", reasoning: true, input: ["text"], contextWindow: 128000, maxTokens: 8192, cost: { input: 0.55, output: 2.19, cacheRead: 0.14, cacheWrite: 0.55 }, enabled: false },
+      { id: "anthropic/claude-sonnet-5", name: "Anthropic: Claude Sonnet 5", reasoning: true, input: ["text", "image"], contextWindow: 1000000, maxTokens: 128000, cost: { input: 2, output: 10, cacheRead: 0.2, cacheWrite: 2.5 }, enabled: false },
+      { id: "anthropic/claude-opus-4.5", name: "Anthropic: Claude Opus 4.5", reasoning: true, input: ["text", "image"], contextWindow: 200000, maxTokens: 64000, cost: { input: 5, output: 25, cacheRead: 0.5, cacheWrite: 6.25 }, enabled: false },
+      { id: "deepseek/deepseek-v4-flash", name: "DeepSeek: DeepSeek V4 Flash", reasoning: true, input: ["text"], contextWindow: 1048576, maxTokens: 393216, cost: { input: 0.14, output: 0.28, cacheRead: 0.028, cacheWrite: 0 }, enabled: false },
     ],
   },
   {
@@ -106,23 +124,27 @@ export const BUILTIN_PROVIDERS: Provider[] = [
     name: "Mistral",
     type: "builtin",
     api: "mistral-conversations",
-    baseUrl: "https://api.mistral.ai/v1",
+    baseUrl: "https://api.mistral.ai",
     hasAuth: false,
     authMethod: "none",
     models: [
-      { id: "mistral-large", name: "Mistral Large", reasoning: false, input: ["text"], contextWindow: 128000, maxTokens: 8192, cost: { input: 2, output: 6, cacheRead: 1, cacheWrite: 2 }, enabled: false },
+      { id: "mistral-large-latest", name: "Mistral Large (latest)", reasoning: false, input: ["text", "image"], contextWindow: 262144, maxTokens: 262144, cost: { input: 0.5, output: 1.5, cacheRead: 0.05, cacheWrite: 0 }, enabled: true },
+      { id: "mistral-medium-latest", name: "Mistral Medium (latest)", reasoning: true, input: ["text", "image"], contextWindow: 262144, maxTokens: 262144, cost: { input: 1.5, output: 7.5, cacheRead: 0.15, cacheWrite: 0 }, enabled: false },
+      { id: "mistral-small-latest", name: "Mistral Small (latest)", reasoning: true, input: ["text", "image"], contextWindow: 256000, maxTokens: 256000, cost: { input: 0.15, output: 0.6, cacheRead: 0.015, cacheWrite: 0 }, enabled: false },
     ],
   },
   {
     id: "github-copilot",
     name: "GitHub Copilot",
     type: "builtin",
-    api: "openai-completions",
-    baseUrl: "https://api.githubcopilot.com",
+    api: "anthropic-messages",
+    baseUrl: "https://api.individual.githubcopilot.com",
     hasAuth: false,
     authMethod: "none",
     models: [
-      { id: "copilot-gpt-4o", name: "Copilot GPT-4o", reasoning: false, input: ["text"], contextWindow: 128000, maxTokens: 4096, cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 }, enabled: false },
+      { id: "claude-sonnet-5", name: "Claude Sonnet 5", reasoning: true, input: ["text", "image"], contextWindow: 1000000, maxTokens: 128000, cost: { input: 2, output: 10, cacheRead: 0.2, cacheWrite: 2.5 }, enabled: true },
+      { id: "gpt-5.6-sol", name: "GPT-5.6 Sol", reasoning: true, input: ["text", "image"], contextWindow: 1050000, maxTokens: 128000, cost: { input: 5, output: 30, cacheRead: 0.5, cacheWrite: 6.25 }, enabled: true },
+      { id: "gemini-3.5-flash", name: "Gemini 3.5 Flash", reasoning: true, input: ["text", "image"], contextWindow: 200000, maxTokens: 64000, cost: { input: 1.5, output: 9, cacheRead: 0.15, cacheWrite: 0 }, enabled: true },
     ],
   },
   {
@@ -134,7 +156,8 @@ export const BUILTIN_PROVIDERS: Provider[] = [
     hasAuth: false,
     authMethod: "none",
     models: [
-      { id: "llama-3.3-70b", name: "Llama 3.3 70B", reasoning: false, input: ["text"], contextWindow: 128000, maxTokens: 4096, cost: { input: 0.59, output: 0.79, cacheRead: 0, cacheWrite: 0 }, enabled: false },
+      { id: "llama-3.3-70b-versatile", name: "Llama 3.3 70B", reasoning: false, input: ["text"], contextWindow: 131072, maxTokens: 32768, cost: { input: 0.59, output: 0.79, cacheRead: 0, cacheWrite: 0 }, enabled: true },
+      { id: "meta-llama/llama-4-scout-17b-16e-instruct", name: "Llama 4 Scout 17B 16E", reasoning: false, input: ["text", "image"], contextWindow: 131072, maxTokens: 8192, cost: { input: 0.11, output: 0.34, cacheRead: 0, cacheWrite: 0 }, enabled: false },
     ],
   },
 ];

package/src/data/model-catalog.test.ts

@@ -0,0 +1,122 @@
+import { describe, expect, it } from "vitest";
+import {
+  MODEL_CATALOG,
+  catalogEntryId,
+  catalogToModel,
+  findCatalogEntry,
+  searchCatalog,
+} from "./model-catalog";
+
+describe("model catalog integrity", () => {
+  it("every entry exposes a callable api id", () => {
+    for (const entry of MODEL_CATALOG) {
+      const id = catalogEntryId(entry);
+      expect(id, `${entry.name} has no id`).toBeTruthy();
+      // A trailing separator means the value is a match prefix, not a model id.
+      expect(id, `${entry.name} -> "${id}" is a prefix, not an id`).not.toMatch(/[-_.:]$/);
+    }
+  });
+
+  it("api ids are unique", () => {
+    const ids = MODEL_CATALOG.map(catalogEntryId);
+    expect(new Set(ids).size).toBe(ids.length);
+  });
+
+  it("picking a preset and re-detecting its id resolves to the same entry", () => {
+    for (const entry of MODEL_CATALOG) {
+      const id = catalogEntryId(entry);
+      expect(findCatalogEntry(id), `${id} round-trips to the wrong entry`).toBe(entry);
+    }
+  });
+
+  it("maxTokens never exceeds contextWindow", () => {
+    for (const entry of MODEL_CATALOG) {
+      expect(entry.maxTokens, `${entry.name}`).toBeLessThanOrEqual(entry.contextWindow);
+    }
+  });
+
+  it("every entry has a cost so usage stats are not silently free", () => {
+    const missing = MODEL_CATALOG.filter((e) => !e.cost).map((e) => e.name);
+    expect(missing).toEqual([]);
+  });
+
+  it("catalogToModel prefills the entry's api id", () => {
+    for (const entry of MODEL_CATALOG) {
+      expect(catalogToModel(entry).id).toBe(catalogEntryId(entry));
+    }
+  });
+
+  it("an empty query lists the whole catalog", () => {
+    expect(searchCatalog("", MODEL_CATALOG.length)).toHaveLength(MODEL_CATALOG.length);
+  });
+});
+
+describe("findCatalogEntry resolution", () => {
+  // Exact ids must beat broader family prefixes.
+  const cases: [string, string][] = [
+    ["claude-opus-4-5-20251101", "Claude Opus 4.5"],
+    ["claude-sonnet-4-5", "Claude Sonnet 4.5"],
+    ["claude-haiku-4-5", "Claude Haiku 4.5"],
+    ["anthropic/claude-sonnet-5", "Claude Sonnet 5"],
+    ["gpt-5", "GPT-5"],
+    ["gpt-5.1", "GPT-5.1"],
+    ["gpt-5.6-sol", "GPT-5.6 Sol"],
+    ["gpt-4o-mini", "GPT-4o Mini"],
+    ["o3", "o3"],
+    ["o3-mini", "o3-mini"],
+    ["o3-pro", "o3-pro"],
+    ["deepseek-v4-flash", "DeepSeek V4 Flash"],
+    ["deepseek-v4-flash-free", "DeepSeek V4 Flash (Free)"],
+    ["deepseek-v4-pro", "DeepSeek V4 Pro"],
+    ["glm-4.5", "GLM-4.5"],
+    ["glm-4.5-air", "GLM-4.5-Air"],
+    ["glm-5", "GLM-5"],
+    ["glm-5.2", "GLM-5.2"],
+    ["kimi-k2.6", "Kimi K2.6"],
+    ["kimi-k2.6-thinking", "Kimi K2.6 Thinking"],
+    ["kimi-k2.7-code", "Kimi K2.7 Code"],
+    ["qwen3-max", "Qwen3 Max"],
+    ["qwen3-coder-plus", "Qwen3 Coder Plus"],
+    ["qwen3-vl-plus", "Qwen3 VL Plus"],
+    ["grok-4.6", "Grok 4.6"],
+    ["gemini-2.5-flash", "Gemini 2.5 Flash"],
+    ["gemini-2.5-pro", "Gemini 2.5 Pro"],
+    ["mistral-large-latest", "Mistral Large 3"],
+  ];
+
+  it.each(cases)("%s resolves to %s", (id, name) => {
+    expect(findCatalogEntry(id)?.name).toBe(name);
+  });
+
+  it("returns undefined for an unknown id", () => {
+    expect(findCatalogEntry("totally-made-up-model-xyz")).toBeUndefined();
+  });
+});
+
+describe("vendor-documented values", () => {
+  // Regression guards for the values that were previously wrong. Each is
+  // sourced from the vendor's own pricing / model docs.
+  const expected: Record<string, { ctx: number; max: number; in: number; out: number }> = {
+    "claude-opus-4-5": { ctx: 200_000, max: 65_536, in: 5, out: 25 },
+    "claude-sonnet-4-5": { ctx: 200_000, max: 65_536, in: 3, out: 15 },
+    "gpt-5.1": { ctx: 400_000, max: 128_000, in: 1.25, out: 10 },
+    "o3-pro": { ctx: 200_000, max: 100_000, in: 20, out: 80 },
+    "gemini-2.5-flash": { ctx: 1_048_576, max: 65_536, in: 0.3, out: 2.5 },
+    "deepseek-v4-pro": { ctx: 1_048_576, max: 131_072, in: 1.32, out: 3.96 },
+    "glm-5.1": { ctx: 204_800, max: 131_072, in: 1.4, out: 4.4 },
+    "qwen3-max": { ctx: 262_144, max: 65_536, in: 1.2, out: 6 },
+  };
+
+  it.each(Object.entries(expected))("%s matches vendor docs", (id, want) => {
+    const entry = findCatalogEntry(id);
+    expect(entry).toBeDefined();
+    expect(entry!.contextWindow).toBe(want.ctx);
+    expect(entry!.maxTokens).toBe(want.max);
+    expect(entry!.cost?.input).toBe(want.in);
+    expect(entry!.cost?.output).toBe(want.out);
+  });
+
+  it("glm-5.2 is a Zhipu model, not SenseNova", () => {
+    expect(findCatalogEntry("glm-5.2")?.family).toBe("GLM");
+  });
+});