> ## Documentation Index
> Fetch the complete documentation index at: https://mintlify.com/cheahjs/free-llm-api-resources/llms.txt
> Use this file to discover all available pages before exploring further.

# Groq

> Lightning-fast inference with free access to multiple models

<Note>
  Groq provides free API access to various open-source models with extremely fast inference speeds powered by their LPU (Language Processing Unit) technology.
</Note>

## Overview

Groq offers free access to multiple language models optimized for their custom LPU hardware, delivering industry-leading inference speeds.

## Rate Limits

Each model has specific rate limits:

| Model Name                          | Requests/Day | Tokens/Minute           |
| ----------------------------------- | ------------ | ----------------------- |
| Allam 2 7B                          | 7,000        | 6,000                   |
| Llama 3.1 8B                        | 14,400       | 6,000                   |
| Llama 3.3 70B                       | 1,000        | 12,000                  |
| Llama 4 Maverick 17B 128E Instruct  | 1,000        | 6,000                   |
| Llama 4 Scout Instruct              | 1,000        | 30,000                  |
| Whisper Large v3                    | 2,000        | 7,200 audio-seconds/min |
| Whisper Large v3 Turbo              | 2,000        | 7,200 audio-seconds/min |
| canopylabs/orpheus-arabic-saudi     | -            | -                       |
| canopylabs/orpheus-v1-english       | -            | -                       |
| groq/compound                       | 250          | 70,000                  |
| groq/compound-mini                  | 250          | 70,000                  |
| meta-llama/llama-guard-4-12b        | 14,400       | 15,000                  |
| meta-llama/llama-prompt-guard-2-22m | -            | -                       |
| meta-llama/llama-prompt-guard-2-86m | -            | -                       |
| moonshotai/kimi-k2-instruct         | 1,000        | 10,000                  |
| moonshotai/kimi-k2-instruct-0905    | 1,000        | 10,000                  |
| openai/gpt-oss-120b                 | 1,000        | 8,000                   |
| openai/gpt-oss-20b                  | 1,000        | 8,000                   |
| openai/gpt-oss-safeguard-20b        | 1,000        | 8,000                   |
| qwen/qwen3-32b                      | 1,000        | 6,000                   |

## Available Models

### Text Generation

<CardGroup cols={2}>
  <Card title="Llama 4 Scout" icon="rocket">
    Latest Llama model with 30K tokens/min
  </Card>

  <Card title="Llama 3.3 70B" icon="meta">
    Powerful 70B parameter model
  </Card>

  <Card title="Groq Compound" icon="brain">
    Groq's proprietary model with 70K tokens/min
  </Card>

  <Card title="Qwen 3 32B" icon="sparkles">
    Efficient multilingual model
  </Card>
</CardGroup>

### Speech Recognition

* **Whisper Large v3** - High-accuracy speech recognition
* **Whisper Large v3 Turbo** - Faster speech recognition

### Safety Models

* **Llama Guard 4 12B** - Content moderation
* **Llama Prompt Guard** - Prompt injection detection

## API Usage

<CodeGroup>
  ```python Python theme={null}
  import openai

  client = openai.OpenAI(
      base_url="https://api.groq.com/openai/v1",
      api_key="YOUR_GROQ_API_KEY"
  )

  response = client.chat.completions.create(
      model="llama-3.3-70b-versatile",
      messages=[
          {"role": "user", "content": "Hello, how are you?"}
      ]
  )

  print(response.choices[0].message.content)
  ```

  ```javascript JavaScript theme={null}
  import OpenAI from 'openai';

  const client = new OpenAI({
      baseURL: 'https://api.groq.com/openai/v1',
      apiKey: process.env.GROQ_API_KEY
  });

  const response = await client.chat.completions.create({
      model: 'llama-3.3-70b-versatile',
      messages: [
          { role: 'user', content: 'Hello, how are you?' }
      ]
  });

  console.log(response.choices[0].message.content);
  ```

  ```bash cURL theme={null}
  curl https://api.groq.com/openai/v1/chat/completions \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer YOUR_GROQ_API_KEY" \
    -d '{
      "model": "llama-3.3-70b-versatile",
      "messages": [
        {"role": "user", "content": "Hello, how are you?"}
      ]
    }'
  ```
</CodeGroup>

## Getting Started

<Steps>
  <Step title="Create Account">
    Sign up at [console.groq.com](https://console.groq.com)
  </Step>

  <Step title="Generate API Key">
    Create an API key from your dashboard
  </Step>

  <Step title="Start Building">
    Use the OpenAI-compatible API for inference
  </Step>
</Steps>

## Key Features

<Info>
  * Fastest inference speeds in the industry
  * OpenAI-compatible API
  * Multiple model options including Llama, Mistral, and Groq models
  * Audio transcription with Whisper
  * Content moderation models
  * Generous free tier limits
</Info>

## Performance Highlights

<CardGroup cols={3}>
  <Card title="Ultra-Fast" icon="bolt">
    Industry-leading inference speeds
  </Card>

  <Card title="High Throughput" icon="gauge-high">
    Up to 70,000 tokens/minute
  </Card>

  <Card title="Low Latency" icon="clock">
    Millisecond response times
  </Card>

  <Card title="LPU Technology" icon="microchip">
    Custom hardware for LLM inference
  </Card>

  <Card title="Multiple Models" icon="layer-group">
    Wide selection of open models
  </Card>

  <Card title="Audio Support" icon="microphone">
    Whisper for speech recognition
  </Card>
</CardGroup>

## Additional Resources

<CardGroup cols={2}>
  <Card title="Groq Console" icon="browser" href="https://console.groq.com">
    Access the platform
  </Card>

  <Card title="Documentation" icon="book" href="https://console.groq.com/docs">
    API documentation
  </Card>
</CardGroup>