Permanent Free LLM API List (API Keys) 😎🔑

This repository curates a comprehensive list of Large Language Model (LLM) APIs offering permanent free tiers, targeting developers and researchers seeking cost-effective access to text inference. It provides a valuable resource for identifying and utilizing LLMs without incurring significant expenses, focusing on sustainable free access rather than temporary trials.

How It Works

The project categorizes LLM API providers into two main groups: "Provider APIs," which are companies that train or fine-tune their own models (e.g., Cohere, Google Gemini, Mistral AI), and "Inference providers," third-party platforms hosting various open-weight models (e.g., Cerebras, Cloudflare Workers AI, Groq, Hugging Face). Each entry details available models, specific rate limits (RPM, RPD, or other metrics), and notes on API compatibility, emphasizing permanent free offerings.

Quick Start & Requirements

This repository is a curated list and does not require installation or execution. To utilize the listed LLM APIs, users must visit the respective provider's website, obtain an API key, and adhere to their specific terms and rate limits. No software prerequisites are needed to browse or reference this list.

Highlighted Details

• Provider APIs: Cohere offers 20 RPM and 1K/mo; Google Gemini provides 5-15 RPM and 100-1K RPD; Mistral AI allows 1 req/s and 1B tok/mo.
• Inference Providers: Cloudflare Workers AI grants 10K neurons/day; Groq offers 30 RPM (1K RPD, or 14,400 RPD for Llama 3.1 8B); Hugging Face provides $0.10/mo in free credits.
• Ollama Cloud measures usage by GPU time, with free tier limits resetting periodically, and is not OpenAI SDK-compatible.
• OpenRouter enables 1K RPD for free models after a $10 credit purchase and includes a free models router.

Maintenance & Community

The project encourages community contributions via pull requests for missing free tiers. Specific details on maintainers, sponsorships, or dedicated community channels (like Discord/Slack) are not provided in the README.

Licensing & Compatibility

The repository's README does not specify a software license. Consequently, users should exercise caution regarding usage rights and compatibility, particularly for commercial applications.

Limitations & Caveats

Some providers have "Limits undocumented." Google Gemini's free tier is unavailable in the EU, UK, or Switzerland. Ollama Cloud's API is not OpenAI SDK-compatible. OpenRouter's enhanced rate limits require a monetary credit purchase. Hugging Face's free offering is credit-based rather than a fixed rate limit.