/01 — Free credits
Free credits & free tiers
Where to get real inference without reaching for a card.
Gemini 2.5 Flash/Pro, Gemini 3.5 Flash — 1M context, 1500 req/day free tier, no credit card required
Llama, Mixtral, Gemma models at record inference speed. Rate-limited free tier, no credit card needed
20 RPM free tier on select models (Llama 3, Gemma, Qwen, Nemotron). Up to 1000 req/day with $10 lifetime topup
GPT-4o, Claude Sonnet, Llama 3, Mistral Large inference inside your GitHub account. Free with GitHub PAT
Serverless Inference API with generous free daily quota. Thousands of community models + Spaces hosting
Mistral Large 3, Medium 3.5, Small 4, Codestral — ~1B tokens/month free tier, no credit card
Llama 3, Nemotron, Mistral, and more on NVIDIA's optimized inference stack. Phone verification required
Free inference for Hermes models, DeepSeek, Qwen, and more. OpenAI-compatible API
Llama, Mistral, Qwen models via Cloudflare's global edge network. 100K requests/day free
Command A, Command R+ models — 1000 API calls/month free tier, non-commercial use. No credit card
AI SDK with built-in support for OpenAI, Anthropic, Google, Mistral. Vercel AI Gateway free tier included
Giga-scale AI inference on wafer-scale chips. Llama and Mistral models at extreme speeds, free tier available
Llama, Mixtral, DeepSeek models with high throughput. $1 in trial credits, no credit card for initial signup
Llama 3, DeepSeek, Qwen models on SN40L chips. Free tier with rate limits, no credit card