Skip to main content
The Vercel AI SDK provider does not expose a built-in listModels() helper. Use one of these patterns to keep model IDs current.

​
Option 1: Fetch from the Gateway API

const response = await fetch("https://api.phaseo.app/v1/models", {
  method: "GET",
  headers: {
    Authorization: `Bearer ${process.env.AI_STATS_API_KEY}`,
  },
});

const filteredResponse = await fetch(
  "https://api.phaseo.app/v1/models?provider=anthropic&status=active,deprecated&availability=all",
  {
    headers: {
      Authorization: `Bearer ${process.env.AI_STATS_API_KEY}`,
    },
  },
);

if (!response.ok || !filteredResponse.ok) {
  throw new Error("Failed to load models");
}

const models = await response.json();
const filteredModels = await filteredResponse.json();
console.log(models, filteredModels);

​
Option 2: Use the official TypeScript SDK

import AIStats from "@ai-stats/sdk";

const client = new AIStats({ apiKey: process.env.AI_STATS_API_KEY! });
const models = await client.getModels({
  provider: ["anthropic"],
  status: ["active", "deprecated"],
  availability: "all",
});

console.log(models);

​
Using discovered IDs with AI SDK

import { aiStats } from "@ai-stats/ai-sdk-provider";
import { generateText } from "ai";

const result = await generateText({
  model: aiStats("openai/gpt-5-nano"),
  prompt: "Hello from a discovered model ID.",
});

console.log(result.text);

​
Notes

  • Model IDs use provider/model format.
  • /v1/models supports filters such as provider, status, organisation, endpoints, and availability.
  • Re-check model availability periodically for routing and deprecations.
  • Prefer selecting IDs from /v1/models over hard-coding long-term.
Last modified on June 10, 2026