Source-backed directory

Compare providers.
Skip the sales pitch.

A neutral starting point for choosing an AI API. We document what can be verified and leave the rest marked as unknown.

Provider typeOpenAI compatibilityPaymentRegionSort

36 providers · reviewed with official sources

0 / 3Compare selected

Add AI21 to comparisonA2AI21AI21's first-party platform for its language models and documented enterprise AI APIs.Provider typeofficialOpenAI compatibilityunknownLast reviewed2026-10-24

Add Alibaba Cloud Model Studio to comparisonALAlibaba Cloud Model StudioAlibaba Cloud's managed platform for Qwen and other supported models, also accessed through DashScope APIs.Provider typeofficialOpenAI compatibilitypartialLast reviewed2026-10-24

Add Anthropic to comparisonANAnthropicFirst-party access to Claude models through the Anthropic API and developer console.Provider typeofficialOpenAI compatibilitypartialLast reviewed2026-10-24

Add AWS Bedrock to comparisonABAWS BedrockAn AWS managed service for accessing foundation models from multiple providers through Bedrock APIs.Provider typeofficialOpenAI compatibilityunknownLast reviewed2026-10-24

Add Azure OpenAI to comparisonAZAzure OpenAIMicrosoft Azure's managed service for accessing OpenAI models with Azure deployment, identity, and networking controls.Provider typeofficialOpenAI compatibilitypartialLast reviewed2026-10-24

Add Cerebras to comparisonCECerebrasA hosted inference service for selected models running on Cerebras infrastructure, with shared and dedicated options.Provider typeofficialOpenAI compatibilitypartialLast reviewed2026-10-24

Add Cloudflare Workers AI to comparisonCFCloudflare Workers AICloudflare's serverless inference platform for running supported models through Workers and REST APIs.Provider typeofficialOpenAI compatibilitypartialLast reviewed2026-10-24

Add Cohere to comparisonCOCohereFirst-party API access to Cohere models for generation, embeddings, reranking, and related language workloads.Provider typeofficialOpenAI compatibilityunknownLast reviewed2026-10-24

Add Databricks to comparisonDBDatabricksDatabricks' managed platform for accessing and serving supported foundation models within a Databricks workspace.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add DeepInfra to comparisonDIDeepInfraA hosted inference platform offering a catalog of language, embedding, image, speech, and other models through developer APIs.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add Featherless to comparisonFEFeatherlessA serverless inference platform providing API access to its supported model catalog.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add Fireworks AI to comparisonFIFireworks AIAn inference and training platform for supported open models, with serverless APIs and dedicated deployment options.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add GitHub Models to comparisonGHGitHub ModelsA GitHub-hosted catalog and API for prototyping with models from multiple providers.Provider typeaggregatorOpenAI compatibilitypartialLast reviewed2026-10-24

Add Google AI Studio to comparisonGOGoogle AI StudioGoogle’s developer entry point for experimenting with and accessing Gemini models.Provider typeofficialOpenAI compatibilitypartialLast reviewed2026-10-24

Add Google Vertex AI to comparisonGVGoogle Vertex AIGoogle Cloud's platform for building with Gemini and other supported models using Google Cloud project controls.Provider typeofficialOpenAI compatibilitypartialLast reviewed2026-10-24

Add Groq to comparisonGRGroqA hosted inference service for supported open-weight models using Groq’s LPU infrastructure.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add Hugging Face Inference Providers to comparisonHFHugging Face Inference ProvidersA unified API for running supported models through Hugging Face and third-party inference providers.Provider typeaggregatorOpenAI compatibilitypartialLast reviewed2026-10-24

Add Jina AI to comparisonJIJina AIJina AI's first-party APIs for embeddings, reranking, reading, and related search-foundation tasks.Provider typeofficialOpenAI compatibilityunknownLast reviewed2026-10-24

Add Lambda to comparisonLALambdaLambda's cloud platform offering hosted inference APIs alongside GPU compute products.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add MiniMax to comparisonMMMiniMaxMiniMax's first-party developer platform for its language, speech, video, and other models.Provider typeofficialOpenAI compatibilityunknownLast reviewed2026-10-24

Add Mistral AI to comparisonMIMistral AIFirst-party access to Mistral models through La Plateforme, with documented model-specific API pricing and capabilities.Provider typeofficialOpenAI compatibilityunknownLast reviewed2026-10-24

Add Moonshot AI to comparisonMOMoonshot AIMoonshot AI's first-party platform for Kimi models and related developer APIs.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add Nebius Token Factory to comparisonNENebius Token FactoryNebius' hosted inference service for supported models, formerly documented as AI Studio.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add Novita AI to comparisonNONovita AINovita AI's hosted platform for supported language, image, and other model APIs.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add NVIDIA NIM to comparisonNVNVIDIA NIMNVIDIA's inference microservices and hosted API catalog for supported models.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add Ollama to comparisonOLOllamaA local model runtime with HTTP APIs and a model library; local inference is not a hosted API provider service.Provider typeotherOpenAI compatibilitypartialLast reviewed2026-10-24

Add OpenAI to comparisonOAOpenAIFirst-party API access to OpenAI models, with usage-based pricing and model-specific limits.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add OpenRouter to comparisonOROpenRouterA unified API and model marketplace that routes requests to multiple upstream providers.Provider typeaggregatorOpenAI compatibilityyesLast reviewed2026-10-24

Add Perplexity to comparisonPPPerplexityPerplexity's API platform for supported Sonar models and search-enabled generation features.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add Replicate to comparisonREReplicateA hosted platform for running public and custom models through Replicate's API.Provider typeofficialOpenAI compatibilityunknownLast reviewed2026-10-24

Add SambaNova to comparisonSNSambaNovaSambaNova's hosted inference platform for supported open-weight and enterprise models.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add Together AI to comparisonTOTogether AIA cloud inference platform offering hosted open-weight models and developer APIs.Provider typeofficialOpenAI compatibilityyesLast reviewed2026-10-24

Add Vercel AI Gateway to comparisonVEVercel AI GatewayA unified gateway for accessing models from multiple providers with Vercel-managed routing and observability.Provider typeaggregatorOpenAI compatibilityyesLast reviewed2026-10-24

Add Voyage AI to comparisonVOVoyage AIVoyage AI's first-party API for embedding and reranking models.Provider typeofficialOpenAI compatibilityunknownLast reviewed2026-10-24

Add xAI to comparisonXAxAIFirst-party developer API access to xAI’s Grok models, with official documentation for models, pricing, and SDK integration.Provider typeofficialOpenAI compatibilitypartialLast reviewed2026-10-24

Add Zhipu AI to comparisonZHZhipu AIZhipu AI's first-party open platform for GLM models and related AI APIs.Provider typeofficialOpenAI compatibilityunknownLast reviewed2026-10-24

What “markup” means here.
A markup compares a resale price with a defined upstream base price using the same model, billing unit, and conditions. These cannot always be established consistently, so we do not publish a universal percentage. “Official” refers to first-party model providers or inference platforms, not an endorsement.