Every top AI model, in one place

Routes every prompt to the best model automatically, and generates an image when you ask for one.

A rotating selection of open models at no token cost.

Fast everyday tasks at no token cost.

Fast, capable and great value for routine work.

A balanced mix of power and cost.

OpenAI's top model for complex reasoning.

The fast, low-cost tier of OpenAI's newest GPT-5.6 family.

The balanced GPT-5.6 — strong performance at mid-tier cost.

OpenAI's newest flagship for the hardest reasoning tasks.

Efficient OpenAI image generation.

High-quality OpenAI image generation.

OpenAI's top-quality image model.

Fast summaries and analysis.

The most popular Claude model — a strong all-rounder.

The newest Sonnet — smarter than 4.6 at a lower price.

Frontier model for long-running autonomous coding and agentic work. 1M context.

Previous-generation Opus, discounted, still highly capable.

Best for nuanced writing and deep reasoning.

Fast and cost-efficient.

Fast multimodal tasks.

Research and long documents.

Advanced reasoning and vision.

Google's capable open-weights model.

Fast, low-cost image generation and editing.

Image generation and faithful editing.

Pro-grade images with 2K/4K detail.

Fast agentic coding model.

Proven, balanced reasoning.

High-speed reasoning model.

xAI's newest flagship model.

Photorealistic image generation.

Fast, lightweight V4 variant.

High performance at low cost.

The renowned open reasoning model, strong at logic and step-by-step problems.

Efficient coding and chat.

Strong coding and analysis.

Mistral's flagship for demanding writing and analysis.

Meta's most capable open model.

Fast, efficient open model for everyday use.

A proven open-source workhorse, free to use.

Fast, low-cost multimodal model that reads images too.

Amazon's capable multimodal model.

Fast and affordable.

The latest Qwen generation.

Qwen's most powerful model, built for demanding reasoning.

Strong long-context reasoning.

Moonshot's newest flagship, with a 1M-token context.

A high-volume open model.

Improved efficiency and faster inference.

Open-weights model, strong at coding and tool use.

Compact but powerful reasoning.

Xiaomi's omnimodal workhorse — near-Pro agentic performance at a fraction of the cost.

A balance of quality and speed across diverse styles.

Top-quality Flux image generation.

Fast, low-cost open Flux model.

A professional design and branding tier.

Photorealistic generation and faithful image-to-image editing.