Best LLM APIs & Model Providers

Claude's API — the pick for long-context work, tool use, and coding agents.

Claude API for developers — long context, careful reasoning.

The most widely supported API surface, and usually the first one a library adds.

GPT-4, embeddings, DALL-E, Whisper — the AI standard API.

One API key across hundreds of models — the cheapest way to compare providers before committing.

One API for every LLM — switch models without code changes.

Absurdly fast inference on open models, which changes what feels possible in a live UI.

Ultra-fast LLM inference on custom LPU silicon.

Strong open-weight models from a European lab, with the licensing that implies.

Open-weight frontier models from a Paris-based AI lab.

Open-weight reasoning and coding models at a fraction of frontier pricing.

Open-weight reasoning and coding models from DeepSeek AI.

Run, fine-tune, and deploy open-source models at scale.

Fast hosted inference for open-source LLMs.

Enterprise-oriented, and the embeddings are the reason most people arrive.

Enterprise-grade LLMs and embedding APIs.

Where open models and datasets live — plus hosted inference when you don't want to run them.

GitHub for machine-learning models and datasets.