A curated list of free LLM inference APIs, covering rate limits, model lists, and special requirements for major platforms like OpenRouter, Google AI Studio, Groq, and Cerebras. Ideal for developers in the prototyping phase.
Overview#
free-llm-api-resources is a Markdown-maintained resource index repository providing developers with a curated list of currently available free or freemium LLM inference services. It is not a code library or SDK, but a high-density information aggregator.
Problems Solved#
- Fragmented Information: LLM providers are numerous, with free policies and quotas scattered across official sites
- Selection Difficulty: Developers struggle to quickly compare rate limits, model availability, and hidden costs
- Compliance Filtering: Explicitly excludes reverse-engineered non-compliant interfaces
Resource Categories#
Completely Free Providers#
- OpenRouter: 20 req/min, 50 req/day; up to 1000 req/day after top-up; Gemma 3, Llama 3.3, Mistral Small, DeepSeek-R1, Qwen
- Google AI Studio: Gemini 2.5 Flash, Gemma 3 series; detailed token/min limits per model
- NVIDIA NIM: Phone verification required; 40 req/min; various open-source models
- Mistral (La Plateforme): Experiment plan requires data training consent; 1 req/s, 500k tokens/min
- HuggingFace Inference: Serverless inference for <10GB models; $0.10/month credit
- Cerebras: Detailed req/min, tokens/min/h/day limits per model
- Groq: req/day and tokens/min per model; Llama, Whisper, Compound series
- Cohere: 20 req/min, 1k req/month; Aya/Command series
- GitHub Models: Limits based on Copilot subscription tier; GPT-4o/o1/o3, Llama, Mistral, DeepSeek
- Cloudflare Workers AI: 10k neurons/day; extensive open-source model list
Trial Credit Providers#
- Fireworks: $1 credit; various open-source models
- Baseten: $30 credit; pay-per-compute-time
- Alibaba Cloud (Intl) Model Studio: 1M tokens per model; Qwen series
- Modal: $5/month on signup, $30/month after payment method绑定
- SambaNova Cloud: $5 credit for 3 months; Llama, Qwen, DeepSeek, Whisper
- Scaleway Generative APIs: 1M tokens free
Key Information Dimensions#
For each listed provider, the project extracts:
- Quota & Rate Limits: Specific values for req/min, req/day, tokens/min
- Model List: Clearly identifies supported free model names
- Special Requirements: Verification, data policy, regional restrictions
Use Cases#
- Technology selection for lowest-cost LLM inference
- Students and researchers utilizing free credits for experiments
- Rapid MVP development with zero initial investment
- Comparing inference speed and quota strategies across platforms