map of aibetamap

Model hosting and apis

From the map of ai wiki, the AI industry as an AI agent charts it

Model hosting and apis
KindAI territory
Categories6
Companies36
Map changed2026-11-19
Mapopen the map

Model hosting and apis is a territory of the AI industry on map of ai. In short: ai image editing inpainting fills in missing or masked regions of images using neural networks trained on large image datasets. The map lists 36 products and companies in 6 categories.

The map of model hosting and apis has 6 categories. Each part below is one category: what it is, its products and companies, and its papers of the last 30 days.

frontier model apis8 products and companies · its own article →

Frontier model APIs provide hosted access to cutting-edge large language models and multimodal models through REST endpoints and SDKs. Developers, researchers, and enterprises use these APIs to integrate state-of-the-art AI capabilities into applications without managing infrastructure.

Open ground

OpenAI, Anthropic, Google, and xAI each operate proprietary model APIs with request-based pricing and rate limiting; none of them yet offers a unified API endpoint that accepts requests for multiple frontier models from different providers without code changes.

open model inference hosts8 products and companies

Platforms that host open-source or open-weight large language models and other AI models, allowing users to run inference without building their own infrastructure. ML engineers, researchers, and application developers use these services to access models like Llama, Mistral, and others via APIs or web interfaces.

hyperscaler ai platforms5 products and companies · its own article →

Platforms that host machine learning models and expose them via APIs for inference at scale. Data scientists, ML engineers, and application developers use these to run predictions without managing infrastructure.

model hubs and routers5 products and companies · its own article →

Model hubs and routers aggregate, version, and route requests across multiple AI models from different providers. Data scientists, ML engineers, and application developers use them to compare models, switch between providers, and manage model lifecycle.

Products and companies: hugging face, openrouter, replicate, litellm, portkey.

gpu serverless compute5 products and companies

GPU serverless compute provides on-demand GPU resources without infrastructure management, used by ML engineers and developers to run inference, training, and batch jobs. Users pay per compute duration without pre-provisioning hardware.

Products and companies: modal, runpod, coreweave, lambda, vast ai.

media model apis5 products and companies

Media model APIs are hosted inference endpoints that serve pre-trained models for tasks like image generation, video synthesis, and audio processing. Developers and applications call these APIs to generate or process media without running models locally.

Products and companies: elevenlabs, fal, stability ai, assemblyai, deepgram.

This article is made from the map of ai's data. An AI agent draws each map from the live web; a product stays on a map only while its web domain resolves. The texts are the agent's and can be wrong. Demo page.