AI Model Comparison

A single-page web app to compare AI models by pricing, capabilities, and benchmarks.

Model Types

• Text/LLM — Language models with token pricing, context lengths, and benchmarks (GPQA, AIME 2025, SWE-bench)
• Image Generation — DALL-E, GPT Image, Imagen with per-image pricing
• Video Generation — Sora, Veo with per-second pricing
• Text-to-Speech — TTS models with voice counts and character pricing
• Realtime Voice — Live audio models with separate text/audio pricing

Features

• Filter by provider, context length, input modalities (image/video/audio/PDF)
• Filter by capabilities (reasoning, open weights)
• Star models to compare side-by-side
• Sort by any column

Quick Start

npm install
npm start        # builds data and serves at http://localhost:8888

Or separately:

npm run build    # generate data/models.json from models.dev submodule
npm run serve    # serve on port 8888

Updating Model Data

To refresh the model data with the latest from upstream:

git submodule update --remote   # pull latest models.dev data
npm run update-models           # regenerate data/models.json

Data Sources

• Model metadata: models.dev (git submodule)
• Benchmarks: llm-stats.com
• Specialized models: OpenAI and Google API docs

Project Structure

index.html       # Main app (vanilla JS + DataTables)
build-data.js    # Generates models.json from models.dev
data/models.json # Generated model data
models.dev/      # Git submodule with model metadata