Every model that matters. On your keys.

AWS's fully managed service for generative-AI apps and agents, offering foundation models from leading providers — including image generation via Amazon Nova, Titan, and Stability AI.

Maker of Claude — a family of frontier AI models known for strong reasoning, long context windows, and reliable, safe responses.

Apertus — the fully open, multilingual LLM from the Swiss AI Initiative (ETH Zurich / EPFL / CSCS), served via the Public AI Inference Utility with a 65k context window.

Microsoft Azure's managed access to OpenAI models, including gpt-image generation — backed by Azure's enterprise security, regional deployment, and compliance controls.

Bitdeer AI Cloud — GPU cloud offering serverless inference and model APIs for open-source and enterprise AI models.

Enterprise AI platform offering the Command language models plus best-in-class embeddings and reranking for search and RAG.

A developer-friendly inference hub serving 100+ models across text, image, video, and speech via simple APIs — pay-as-you-go on its own cost-optimized US infrastructure.

AI lab behind the DeepSeek open models, offering strong reasoning and coding performance at very low cost.

A high-performance inference platform serving frontier open models — including FLUX image generation — at open-source economics, processing 30T+ tokens per day.

GMI Cloud — an NVIDIA-powered inference cloud that serves open models through a unified API.

Google's Gemini models via the Gemini Developer API — multimodal AI for text, images, and video using your Google AI Studio key.

Google Cloud's unified AI platform for building and scaling generative apps — featuring Imagen for photorealistic images and Veo for high-quality text-to-video, with governance built in.

Groq runs AI inference on its purpose-built LPU chip for ultra-low latency. For speech it serves Whisper ASR and Orpheus/PlayAI TTS — real-time transcription and voice generation.

Lightning AI — a platform to build, train, and deploy AI models and agents, with hosted model APIs.

Meta's open-weight Llama models — freely available large language models you can self-host or use via hosting partners.

Hailuo, MiniMax's AI video generator, is known for striking cinematic motion and strong prompt following from text or a single image — with templates for dance, effects, and character animation.

European AI lab offering the Mistral family of efficient open and commercial language models.

Maker of Kimi — Moonshot AI's frontier models (Kimi K2) with long context windows, strong reasoning, and an OpenAI-compatible API.

Maker of Nemotron — NVIDIA's open models for agentic reasoning and tool calling, served through the hosted NIM catalog with an OpenAI-compatible API.

OpenAI's image generation via gpt-image-1 — the natively multimodal model behind ChatGPT — delivering versatile styles, strong world knowledge, accurate text, and prompt-driven edits.

Bring any OpenAI Chat Completions–compatible endpoint — hosted gateways, inference platforms, or your own deployment — by supplying its Base URL and API key.

A unified gateway to 400+ models from 70+ providers through a single OpenAI-compatible API — including a dedicated image API spanning 30+ models — with automatic failover.

AI answer engine; its Sonar API returns up-to-date, web-grounded responses with citations.

Alibaba's Qwen family, served via DashScope / Model Studio — Qwen-Image generates native 2K images from long prompts, and Wan delivers text-to-video and image-to-video.

A lightning-fast inference platform serving 200+ open and commercial models — LLMs plus image, video, and audio — through a single OpenAI-compatible API with predictable pricing.

A full-stack inference cloud serving 200+ open-source models through one API — chat, image, audio, and video — with optimized kernels for faster, cheaper generation at scale.

Vercel AI Gateway routes requests to hundreds of models across many providers through a single API — text, image, and video — with automatic failover and no platform markup.

Vultr — cloud provider offering serverless inference for open models alongside S3-compatible object storage.

xAI builds Grok, the AI assistant from Elon Musk's xAI with real-time knowledge of the world. Its image model (Aurora) renders photorealistic, prompt-faithful images directly from the same API key as the Grok chat models.

Maker of GLM — Z.AI's (Zhipu) frontier models with agentic reasoning, long context, and multimodal input, served through an OpenAI-compatible API.