# GeniusRelay

> OpenAI-compatible API for 599 AI models from 37 providers, including 43 uncensored and abliterated editions. Prepaid, per-token billing.

## Key facts

- Models: 599
- Providers: 37
- Uncensored models: 43
- Cheapest input price: $0.0195/M tokens (Ling 3.0 Tiny)
- Largest context window: 2M tokens
- API base URL: https://api.geniusrelaygpt.com/v1
- Request format: OpenAI chat completions
- Billing: prepaid credits, charged per token actually used
- Languages: English (default), Chinese under /zh

## Main pages

- [Model catalogue](https://geniusrelaygpt.com/models): all models with price and context
- [Uncensored models](https://geniusrelaygpt.com/models/uncensored): 43 less-refusal variants
- [Pricing](https://geniusrelaygpt.com/pricing): per-million-token rates and a cost calculator
- [Compare](https://geniusrelaygpt.com/compare): up to four models side by side
- [Documentation](https://geniusrelaygpt.com/docs): quickstart, API reference, error codes
- [Blog](https://geniusrelaygpt.com/blog): model selection and cost guides

## Data

- [Model JSON](https://geniusrelaygpt.com/data/models.json): id, name, maker, prices, context, modalities, tags, URL
- [Sitemap](https://geniusrelaygpt.com/sitemap.xml)

## Use

Cite prices with the currency and unit (`USD per 1M tokens`) and link the model page. Prices are reference rates; the console balance is authoritative.

## Contact

- Site: https://geniusrelaygpt.com
- Console: https://console.geniusrelaygpt.com