Free Groq API Key, Base URL & Rate Limits
API ProviderGroq is a cloud AI platform powered by its proprietary LPU (Language Processing Unit) chips, delivering dramatically faster inference than GPU-based providers.
How to get a free Groq API key
- 1
- 2 Go to API Keys in the sidebar
- 3 Create API key
- 4 Choose a model Llama 3.3 70B is the most popular free option.
- 5 Configure OpenAI client Base URL: https://api.groq.com/openai/v1
Provider Snapshot
Supported Models 12 models
View in directory →| Model ID | Developer | Context | Availability | Free Tier | Use cases |
|---|---|---|---|---|---|
| moonshotai/kimi-k2-instruct Moonshot Kimi K2 | Kimi | 131K | Online | Yes | chatcoding |
| moonshotai/kimi-k2-instruct-0905 Moonshot Kimi K2 0905 | Kimi | 131K | Online | Yes | chatcoding |
| groq/compound groq/compound | Groq | 131K | Online | Yes | chat |
| groq/compound-mini groq/compound-mini | Groq | 131K | Online | Yes | chat |
| whisper-large-v3 whisper-large-v3 | Groq | 131K | Online | Yes | chat |
| whisper-large-v3-turbo whisper-large-v3-turbo | Groq | 131K | Online | Yes | chat |
| allam-2-7b allam-2-7b | Groq | 8K | Online | Yes | chat |
| deepseek-r1-distill-70b deepseek-r1-distill-70b | Groq | 131K | Check provider | Yes | chatreasoning |
| kimi-k2-instruct kimi-k2-instruct | Kimi | 262K | Check provider | Yes | chat |
| llama-4-maverick-17b-128e-instruct llama-4-maverick-17b-128e-instruct | Groq | 131K | Check provider | Yes | chat |
| meta-llama/llama-4-scout-17b-16e-instruct llama-4-scout-17b-16e-instruct | Groq | 131K | Check provider | Yes | chat |
| qwen/qwen3-32b qwen3-32b | Alibaba | 131K | Check provider | Yes | chat |
Developer Tools
Related Providers
View all →Groq FreeLLM Score free API access score
How we score →What is Groq?
World's fastest LLM inference — ultra-low latency, free tier.
Groq is a cloud AI platform powered by its proprietary LPU (Language Processing Unit) chips, delivering dramatically faster inference than GPU-based providers. The free tier supports Llama, Qwen, DeepSeek-R1, and Whisper models with generous daily limits. Groq is fully OpenAI SDK-compatible, making it a drop-in replacement for any tool that accepts a custom base URL.
- Ultra-fast inference (~2,600 tok/s)
- Free tier: 14,400 RPD for most models
- Supports Llama 4, Qwen3, DeepSeek-R1
- OpenAI-compatible
API Compatibility: Live-tested OpenAI-compatible Chat Completions and Responses
Groq Free Tier Limits & Pricing
Groq API Setup Tutorials
Groq is fully compatible with popular AI coding assistants like Cursor, Claude Code, and more. To see step-by-step API configuration instructions for your favorite tool, please visit our Global Configuration Guide →
Groq Model Use Cases
What Groq's free models are best for, based on aggregated model capabilities:
Groq Limitations & Caveats
- Rate limits vary significantly by model — check per-model limits
- Some models have token-per-minute caps in addition to RPM
- LPU availability may cause queuing during peak usage
Groq FAQ
Why are Groq's rate limits different for each model?
Groq's LPU hardware has model-specific throughput. Larger models (70B+) get lower RPM, while smaller models (8B) can handle 30 RPM or more. Always check the per-model rate card in the Groq console.
Is Groq really faster than other free LLM providers?
Yes — Groq's LPU chips deliver 2,000-3,000 tokens/second on smaller models, which is 5-10× faster than GPU-based providers. This makes Groq ideal for real-time applications like chatbots and coding assistants.
Can I use Groq as a drop-in replacement for OpenAI?
Yes. Groq's API is fully OpenAI-compatible. Just change the base URL to https://api.groq.com/openai/v1 and use your Groq API key. Model names differ (e.g. llama-3.3-70b-versatile instead of gpt-4o).