vantage.sh/models

Comparing AI model pricing and benchmarks across cloud vendors shouldn't require dozens of browser tabs.

vantage.sh/models displays AI model pricing, benchmarks, and metadata from various cloud vendors in a single, queryable interface. Built with Astro, React, and Tailwind CSS, it features a client-side SQL engine (sql.js) for querying model data.

Project status

Vantage employees are actively maintaining and hosting the site with the help of contributors here. Improvements in the form of pull requests or ideas via issues are welcome!

We also have a Slack Community for anyone to join.

Requirements

• Node.js 24+

Developing locally

First, install dependencies:

npm install

You then need to either:

• Scrape the data locally to generate public/data.json
• Or obtain a pre-built data.json from CI artifacts if you aren't modifying the scraper

Once you have the data, start the development server:

npm run dev

This starts the local dev server at localhost:4321.

Before making a pull request, run the type checker and formatter:

npm run typecheck
npm run format:check

Scraping the data

You'll need to provide credentials so that the scrapers can access AWS APIs. Options for setting this up are described in the AWS SDK docs.

Once credentials are configured, run:

npm run init

This executes the scrapers in scraper/ and outputs model data to public/data.json.

Adding new vendors

1. Create a new scraper in scraper/scrapers/
2. Add the scraper call to scraper/runner.ts
3. Update the PROVIDERS map as needed
4. Update MODEL_REASONING_PREFIXES and isSelfHostableModel() in scraper/constants.ts for new model families

Commands

All commands are run from the root of the project:

Command	Action
npm install	Install dependencies
npm run dev	Start local dev server at localhost:4321
npm run build	Build for production to ./dist/
npm run preview	Preview the production build locally
npm run init	Run scrapers to generate public/data.json
npm run typecheck	Run TypeScript type checking
npm run format	Format code with Prettier

Architecture

Data pipeline

1. Scrapers (scraper/) fetch model data from vendor APIs (currently AWS Bedrock)
2. Scrapers output to public/data.json following the DataFormat type
3. At build time, src/pages/data.db.ts converts JSON to a SQLite database
4. Client-side sql.js workers query the database for table display

Key files

• src/dataFormat.d.ts - Core type definitions for vendors and models
• src/sql/schema.ts - SQLite schema definition
• scraper/constants.ts - Model metadata (reasoning capability, self-hostable status, benchmark data)
• src/constants.ts - Default table column queries
• src/state.ts - Client-side state management with localStorage persistence
• src/sqlEngine.ts - Web worker pool for SQL queries

Path alias

Use @/src/* to import from the src/ directory (configured in tsconfig.json).

API Access

The data backing vantage.sh/models is available via /models/data.db.

Keep up-to-date

Feel free to watch/star this repo as we're looking to update the site regularly. Vantage also works on the following relevant projects:

• EC2Instances.info - A comparison of EC2 instance types and pricing.
• cur.vantage.sh - Definitions of all AWS Cost and Usage Report (CUR) billing codes by service.
• The Cloud Cost Handbook - An open-source set of guides for best practices of managing cloud costs.
• The AWS Cost Leaderboard - A hosted site of the top AWS cost centers.
• Vantage - A cloud cost transparency platform.