Ad 320×100
AI Directory

Free AI Models & Tokens

Directory of free AI models, GPU computing, and API tokens — use top LLMs without paying

Hugging Face ZeroGPU
Free GPU
Dynamic GPU allocation for Spaces — get free NVIDIA H200 access for ML inference
Free Quota: Free: 5 min/day, PRO ($9/mo): 40 min/day
Features: NVIDIA H200 (70GB VRAM), Gradio SDK, PyTorch compatible, auto-release GPU
Token-Free Gateway
Token-Free Gateway
OpenAI-compatible API using web sessions — no API keys needed, 13 providers supported
Free Quota: Unlimited (browser login based)
Features: ChatGPT, Claude, Gemini, DeepSeek, Grok + 8 more, full function calling
OpenClaw Zero Token
Token-Free Gateway
Use all major AI models without API tokens — browser login, unified gateway
Free Quota: Unlimited (web session based)
Features: DeepSeek, Qwen, Claude, ChatGPT, Gemini, Grok, Kimi, Doubao, Manus
China Telecom Xirang Cloud
Free API
25 million free tokens with phone registration — OpenAI-compatible API
Free Quota: 25M tokens free (2-week validity)
Features: DeepSeek-V3, GLM-5, Doubao-Seed-2.0-pro, OpenAI-compatible endpoint
Google AI Studio
Free API
Free access to Gemini models with generous rate limits for prototyping
Free Quota: Free tier: 60 requests/min
Features: Gemini Pro, Gemini Flash, multimodal, function calling, grounding
Groq
Free API
Ultra-fast LLM inference with free tier — powered by custom LPU hardware
Free Quota: Free: 30 req/min, 14.4K tokens/min
Features: Llama 3, Mixtral, Gemma, ultra-low latency, OpenAI-compatible API
Mistral AI
Free API
Free tier for experimenting with Mistral models via La Plateforme
Free Quota: Free: 1 req/sec, 500K tokens/min
Features: Mistral Large, Mistral Small, Codestral, function calling, JSON mode
Together AI
Free API
Free credits for open-source model inference — fast serverless endpoints
Free Quota: $5 free credits on signup
Features: Llama 3, Mixtral, DBRX, CodeLlama, serverless & dedicated, fine-tuning
OpenRouter
Free API
Unified API for 100+ models — some with free tiers, pay only for what you use
Free Quota: Free models available (rate-limited)
Features: GPT-4, Claude, Llama, Gemini, unified API, automatic fallback routing
SiliconFlow
Free API
Free inference for popular open-source models — fast cloud GPU platform
Free Quota: Free tier for select models
Features: DeepSeek, Qwen, Llama, ChatGLM, image generation, OpenAI-compatible
DeepSeek Platform
Free API
OpenAI-compatible API with strong reasoning models — free credits for new accounts
Free Quota: 5M free tokens on signup (no credit card)
Features: DeepSeek V3/V4, R1 reasoning, chat.deepseek.com free chat, very low paid rates
Cerebras Cloud
Free API
Ultra-fast inference on custom wafer-scale chips — one of the fastest free APIs
Free Quota: 1M tokens/day free, 30 req/min, no credit card
Features: Llama 3.3 70B, Qwen3 235B, GPT-OSS 120B, ~2,000+ tokens/sec
SambaNova Cloud
Free API
High-performance inference with large models on custom RDU hardware
Free Quota: $5 free credits for 3 months on signup
Features: Llama 3.1 405B, DeepSeek R1, Qwen, Gemma — largest models on free tier
Cloudflare Workers AI
Free API
Run LLMs at the edge on Cloudflare's global network — ideal for web apps
Free Quota: 10,000 neurons/day free
Features: Llama 3.3 70B, Kimi K2, GLM-4.7, GPT-OSS, Granite — edge deployment
GitHub Models
Free API
Access frontier and open models through GitHub — great for developers already on GitHub
Free Quota: ~50 chat + 2K completions/month (varies by Copilot tier)
Features: GPT-4o, GPT-4.1, Llama 3.3, DeepSeek, Mistral, Phi-4, playground + API
NVIDIA NIM
Free API
NVIDIA inference microservices with a broad model catalog for evaluation
Free Quota: 40 requests/min free tier
Features: 46+ models, Llama, Mistral, Nemotron, OpenAI-compatible, no credit card
Hugging Face Inference
Free API
Serverless inference for thousands of open models on Hugging Face Hub
Free Quota: Free credits for Inference API / Inference Providers
Features: DeepSeek, Llama, Qwen, Flux image models, embeddings, serverless endpoints
Cohere
Free API
Enterprise-grade NLP with Command models — trial tier for experimentation
Free Quota: 1,000 requests/month on trial
Features: Command R+, Command A, embeddings, rerank, multilingual, function calling
Zhipu AI (GLM)
Free API
China's leading GLM model family with free tier for developers
Free Quota: Free tier with monthly token allowance
Features: GLM-4.5, GLM-4.7 Flash, GLM-Z1, image generation, OpenAI-compatible API
Moonshot Kimi
Free API
Long-context Kimi models with generous free tier for Chinese and English
Free Quota: Free tier with rate limits on signup
Features: Kimi K2, Moonshot v1 8K/32K/128K, 200K+ context, OpenAI-compatible
Ollama
Token-Free Gateway
Run open-weight LLMs locally — completely free, private, no API keys
Free Quota: Unlimited (your own hardware)
Features: DeepSeek R1/V4, Llama 3, Qwen, Mistral, Gemma — OpenAI-compatible local server
FreeLLMAPI
Token-Free Gateway
Self-hosted proxy aggregating 14 free provider tiers into one OpenAI endpoint
Free Quota: ~800M+ tokens/month combined across providers
Features: Gemini, Groq, Cerebras, Mistral, GitHub Models, Cloudflare — auto failover
Fireworks AI
Free API
Fast inference platform with free starter credits for open and proprietary models
Free Quota: $5–6 free credits on signup
Features: DeepSeek, Llama, Mixtral, fine-tuning, serverless, OpenAI-compatible API
Replicate
Free API
Run open-source ML models in the cloud — pay-per-second with free credits
Free Quota: Free credits for new accounts
Features: Llama, Flux, SDXL, Whisper, custom models, REST API, no GPU setup
Alibaba DashScope
Free API
Alibaba Cloud's Qwen model platform with free tier for developers
Free Quota: Free tier with monthly token quota
Features: Qwen 3.5 Plus/Turbo, Qwen-VL vision, embeddings, OpenAI-compatible API
Baidu Qianfan
Free API
Baidu's ERNIE and open model platform — free trial for Chinese developers
Free Quota: Free trial credits on registration
Features: ERNIE 4.0, DeepSeek, Llama, embedding models, OpenAI-compatible endpoint
MiniMax
Free API
Chinese AI lab with abab and hailuo models — free tier for API access
Free Quota: Free tier with limited monthly usage
Features: abab 6.5, MiniMax-Text-01, speech synthesis, video generation API
Scaleway Generative APIs
Free API
European cloud provider with permanent free token allowance
Free Quota: 1M tokens/month free (permanent)
Features: DeepSeek, Llama, Mistral, Gemma — GDPR-friendly EU hosting
Hyperbolic
Free API
GPU cloud with free trial credits for fast open-model inference
Free Quota: $1 free trial credits on signup
Features: DeepSeek, Llama, Qwen, GPT-OSS — competitive pricing after trial
AI21 Labs
Free API
Jurassic and Jamba models with a free trial for developers
Free Quota: Free trial tier with limited requests
Features: Jamba 1.5, Jurassic-2, long context, summarization, structured output

Information may change. Always check the provider's official site for the latest details.

Free AI Models & Tokens

Browse a free, up-to-date directory of AI models you can use without paying. Compare open-weight and free-tier large language models by provider, context window, and capabilities to find the best free option for chat, coding, or experimentation.

FAQ

What does the free AI models directory include?
It lists notable free-to-use and open-weight models with key details like the provider, context length, and what each model is best suited for.
Are these models really free?
The directory focuses on models with a free tier or open weights. Always check each provider's current terms, since limits and pricing can change.
How do I choose the right model?
Match the model to your task: larger context windows suit long documents, while coding-tuned models are better for development. Compare the listed capabilities to decide.