Free LLM API Keys
& Free-Tier AI APIs
A curated directory of 21 providers giving away free, OpenAI-compatible access to Gemini, Llama, DeepSeek, Qwen, Mistral and GPT-4o. Create your own key — no shady shared keys, no surprise bills.
All free LLM API providers
Search and filter 21 providers. Click any card for setup, limits and code.
Showing 21 of 21 providers
Google AI Studio (Gemini)
The most generous permanent free tier for a frontier model.
Groq
The fastest free inference on the planet — 500+ tokens/sec.
OpenRouter
One key, 300+ models — dozens are completely free.
Cerebras
Wafer-scale inference — frontier open models at record speed, free.
GitHub Models
Free access to GPT-4o, o-series and more — just a GitHub account.
Mistral AI
European frontier models with a free “Experiment” tier.
Cloudflare Workers AI
Serverless inference with a free daily allocation on the edge.
SambaNova Cloud
High-speed inference for large open models, free to start.
NVIDIA NIM
Free credits across a huge catalog of hosted open models.
Together AI
Free signup credit plus a free Llama endpoint.
Alibaba Qwen (DashScope)
First-party Qwen models with a free token quota.
Hugging Face
Monthly free inference credits across many providers.
Hyperbolic
Low-cost open-model inference with free starter credit.
Zhipu AI (GLM)
Free GLM-4 Flash — a capable always-free chat model.
Cohere
Free trial keys for Command models, embeddings and rerank.
Fireworks AI
Fast open-model serving with free starter credit.
DeepSeek
Frontier reasoning at rock-bottom prices (free starter credit).
Pollinations AI
Text and image generation with no key and no signup.
xAI (Grok)
Grok models with recurring free API credits.
Scaleway Generative APIs
EU-hosted open models with a free beta allowance.
OVHcloud AI Endpoints
European sovereign-cloud inference, free in beta.
Free API by use case
Pick the best free API for your project.
- Best free LLM API for coding
- Best free LLM API for a chatbot
- Best free embedding & LLM API for RAG
- Best free multimodal / vision API
- Best free speech-to-text API
- Best free LLM API for AI agents
- Best free LLM API for a Telegram bot
- Best free LLM API for a Discord bot
- Best free LLM API for data analysis
- Best free LLM API with no credit card
- Best free GPT-4 alternative API
Free by model
Which providers host each model free.
- Free Llama 3.3 70B API
- Free Gemini 2.5 Flash API
- Free DeepSeek R1 API
- Free GPT-4o API
- Free DeepSeek V3 API
- Free Gemini 2.5 Pro API
- Free Llama 4 Maverick API
- Free Llama 3.1 8B API
- Free Qwen 3 235B API
- Free Qwen 3 Coder API
- Free Whisper Large v3 API
- Free Mistral Small 3.2 API
Popular comparisons
Head-to-head free API breakdowns.
- Google AI Studio (Gemini) vs Groq
- Groq vs Cerebras
- Google AI Studio (Gemini) vs OpenRouter
- OpenRouter vs GitHub Models
- Google AI Studio (Gemini) vs GitHub Models
- Groq vs OpenRouter
- DeepSeek vs Google AI Studio (Gemini)
- Groq vs SambaNova Cloud
- Cerebras vs SambaNova Cloud
- OpenRouter vs Hugging Face
- Mistral AI vs Google AI Studio (Gemini)
- NVIDIA NIM vs Together AI
How free LLM APIs work
Nearly every provider in this directory is OpenAI-compatible. That means you keep your existing code and change only two things: the base_url and the API key you created on the provider’s own site. The same snippet below works for Gemini, Groq, Cerebras, OpenRouter and more — just swap the base URL.
from openai import OpenAI
# Works with almost every provider in this directory —
# just change base_url + the free key you created.
client = OpenAI(
base_url="https://generativelanguage.googleapis.com/v1beta/openai/",
api_key="YOUR_FREE_KEY"
)
resp = client.chat.completions.create(
model="gemini-2.5-flash",
messages=[{"role": "user", "content": "Hello!"}]
)
print(resp.choices[0].message.content)1. Pick a provider
Choose any card above. Free tiers (Gemini, Groq) reset daily; free credits (NVIDIA, Together) give a balance to spend.
2. Create your own key
Click through to the provider’s official key page and generate a key. It is yours — we never see or store it.
3. Swap the base URL
Point the OpenAI SDK at the provider’s base URL (listed on each provider page) and paste your key.
4. Call a model
Use any model the provider offers. Need more scale? Move to their cheap paid tier later with zero code changes.