models.do

Models.do - Search, Filter, and Compare AI Models

Models.do enables developers to search, filter, and compare Large Language Models based on their capabilities and attributes.

Installation

npm install models.do
# or
yarn add models.do
# or
pnpm add models.do

Usage

import { ModelsSDK } from 'models.do'

// Create the SDK client
const models = new ModelsSDK({
  apiKey: 'your-api-key',
})

// Resolve a specific model
await models.get('gemini') // -> { model: ModelDetails, parsed: { model: 'gemini' } }
await models.get('gemini(seed:1)') // -> { model: ModelDetails, parsed: { model: 'gemini', systemConfig: { seed: 1 } } }
await models.get('claude-3.7:reasoning') // -> { model: ModelDetails, parsed: { model: 'claude', capabilities: ['reasoning'] } }

// Find models by filters and sorting
// For most use cases, you'll most likely use the find method as its more flexible
await models.find({
  author: 'google',
  capabilities: ['reasoning', 'structuredOutput'],
  sortBy: 'pricingLowToHigh',
})

// Returns both model's details, as well as a diff of the capabilities and attributes
await models.compare(['gemini', 'gpt-4o'])

// Retrieve a group of pre-parsed models
const frontierModels: Model[] = await models.group('frontier')
const wideRange: Model[] = await models.group('wideRange')

API Reference

ModelsSDK

The main client for interacting with Models.do. Fetches data using the models API provided by Drivly.ai, making the SDK both slim and fast.

Constructor

new ModelsSDK(options?: { apiKey?: string, baseUrl?: 'https://models.do' | string })

Methods

• get(modelIdentifier: string): Promise<{ model: ModelDetails }>
  Get details about a specific model
• find(filters?: ModelFilters): Promise<Record<string, ModelDetails>>
  Find models that match the filters
• compare(modelIdentifiers: string[]): Promise<Record<string, ModelDetails>>
  Compare models based on their capabilities
• group(groupName: string): Promise<Model[]>
  Retrieve a group of models pre-parsed

Types

type Capability = 'code' | 'online' | 'reasoning' | 'tools' | 'structuredOutput' | 'responseFormat'

type Model = {
  isComposite?: boolean
  name: string
  author: string
  modelIdentifier?: string
  openRouterSlug?: string
  provider: Provider
  capabilities?: Capability[]
  alias?: string
  sorting: {
    topWeekly: number
    newest: number
    throughputHighToLow: number
    latencyLowToHigh: number
    pricingLowToHigh: number
    pricingHighToLow: number
  }
}

type ModelFilters = {
  provider?: string
  author?: string
  sortBy?: 'topWeekly' | 'newest' | 'throughputHighToLow' | 'latencyLowToHigh' | 'pricingLowToHigh' | 'pricingHighToLow'
  capabilities?: Capability | Capability[]
  // Shortcuts for different capabilities
  // For example, instead of writing `capabilities: ['structuredOutput', 'tools']`
  // you can write `outputType: 'Object'`
  outputType?: 'Object' | 'ObjectArray' | 'Text' | 'TextArray' | 'Markdown' | 'Code'
}

Related packages

• functions.do - Typesafe AI Functions
• workflows.do - Business Process Automation
• agents.do - Autonomous Digital Workers
• apis.do - Clickable Developer Experiences
• llm.do - Intelligent AI Gateway

License

MIT

Dependencies

• apis.do - Unified API Gateway for all domains and services in the .do ecosystem