language-models

Stop memorizing model IDs. Start shipping.

You're building AI-powered applications, but every provider has different naming conventions. Is it claude-opus-4-5-20251101 or anthropic/claude-opus-4.5? Was it gpt-4o or openai/gpt-4o? You shouldn't have to care.

The Problem

// Without language-models: fragile, provider-specific, constantly breaking
const model = 'anthropic/claude-3-opus-20240229' // Wait, is this still current?
const model = 'claude-opus-4-5-20251101'         // Or was it this format?
const model = 'anthropic/claude-opus-4.5'        // Which one does OpenRouter want?

The Solution

// With language-models: simple, memorable, always resolves correctly
import { resolve } from 'language-models'

resolve('opus')   // 'anthropic/claude-opus-4.5'
resolve('sonnet') // 'anthropic/claude-sonnet-4.5'
resolve('gpt')    // 'openai/gpt-4o'
resolve('llama')  // 'meta-llama/llama-4-maverick'

Quick Start

1. Install

npm install language-models

2. Import

import { resolve, list, search } from 'language-models'

3. Use

// Resolve human-friendly aliases to full model IDs
const modelId = resolve('opus')

// Search across 200+ models
const claudeModels = search('claude')

// Get full model catalog with pricing and context info
const allModels = list()

API Reference

`resolve(input: string): string`

Resolve an alias or partial name to a full OpenRouter model ID.

resolve('opus')                      // 'anthropic/claude-opus-4.5'
resolve('sonnet')                    // 'anthropic/claude-sonnet-4.5'
resolve('gpt')                       // 'openai/gpt-4o'
resolve('llama')                     // 'meta-llama/llama-4-maverick'
resolve('anthropic/claude-opus-4.5') // Pass-through for full IDs

`resolveWithProvider(input: string): ResolvedModel`

Get full routing information including provider details for direct SDK access.

const info = resolveWithProvider('opus')
// {
//   id: 'anthropic/claude-opus-4.5',
//   provider: 'anthropic',
//   providerModelId: 'claude-opus-4-5-20251101',
//   supportsDirectRouting: true,
//   model: { name, pricing, context_length, ... }
// }

`list(): ModelInfo[]`

Get the complete model catalog with pricing, context lengths, and capabilities.

`get(id: string): ModelInfo | undefined`

Fetch a specific model by exact ID.

`search(query: string): ModelInfo[]`

Find models matching a search query across IDs and names.

Supported Aliases

You type	You get
`opus`	`anthropic/claude-opus-4.5`
`sonnet`	`anthropic/claude-sonnet-4.5`
`haiku`	`anthropic/claude-haiku-4.5`
`gpt`, `gpt-4o`, `4o`	`openai/gpt-4o`
`o1`, `o3`, `o3-mini`	`openai/o1`, `openai/o3`, `openai/o3-mini`
`gemini`, `flash`	`google/gemini-2.5-flash`
`gemini-pro`	`google/gemini-2.5-pro`
`llama`, `llama-4`	`meta-llama/llama-4-maverick`
`llama-70b`	`meta-llama/llama-3.3-70b-instruct`
`mistral`	`mistralai/mistral-large-2411`
`codestral`	`mistralai/codestral-2501`
`deepseek`	`deepseek/deepseek-chat`
`r1`	`deepseek/deepseek-r1`
`qwen`	`qwen/qwen3-235b-a22b`
`grok`	`x-ai/grok-3`
`sonar`	`perplexity/sonar-pro`

Direct Provider Routing

For providers that support direct SDK access (Anthropic, OpenAI, Google), use resolveWithProvider to get the native model ID:

import { resolveWithProvider, DIRECT_PROVIDERS } from 'language-models'

const { provider, providerModelId, supportsDirectRouting } = resolveWithProvider('opus')

if (supportsDirectRouting) {
  // Use native SDK with providerModelId
} else {
  // Route through OpenRouter
}

Updating the Model Catalog

pnpm fetch-models

Fetches the latest models from OpenRouter and updates data/models.json.

License

MIT