APIMODELS

APIMODELS is a pay-as-you-go gateway that exposes 150+ image, video, chat and audio models through one API key, with OpenAI and Anthropic SDK compatibility, MCP support and prices listed below official rates.

APIMODELS is an aggregation layer for AI model APIs. Instead of opening separate accounts with Anthropic, OpenAI, Google, ByteDance, ElevenLabs and others, a developer creates one key and calls image, video, language and audio models through a single endpoint. The catalog covers more than 150 models, including Claude Opus 5.5, GPT-6 Sol, Gemini, Wan 3.0, MiniMax H3, Kling, the Seedance line, Nano Banana, GPT Image and DeepSeek, Qwen and GLM chat models. The service is aimed at independent developers, small teams and creators who want frontier models without large minimum deposits or organization verification. Billing is pay as you go with no monthly fee, top-ups start at $10, and failed requests are not charged. The homepage lists prices below official rates for many models, from roughly 10% off for DeepSeek V4.1 Flash to about 63% off for some image models, with a claim of up to 95% savings overall. Payment works through Stripe, PayPal, Alipay and USDT, and the endpoint is described as reachable from mainland China without a proxy. In practice, integration is mostly a base URL change. The OpenAI and Anthropic SDKs work by pointing base_url at the APIMODELS endpoint and swapping the model string, with chat completions and Anthropic-style messages both supported. Popular models run on several upstream channels with automatic failover, long video jobs run asynchronously with webhook callbacks, and any request can be looked up by record ID. An MCP server lets Claude Desktop, Cursor and Cherry Studio generate images, video, speech and chat with the same key. Each model has its own page with live pricing, and the site also publishes prompt libraries for several video and image models. Among alternatives, it sits between going direct to each provider and using other multi-model gateways. The trade-off is that capabilities and rate limits are inherited from upstream providers, generated files expire after 30 days, and there is no permanent free tier beyond a small trial credit.