groq-models.yaml

provider: groq
name: Groq
links:
  - name: Home
    link: https://groq.com
  - name: Models
    link: https://console.groq.com/docs/models
  - name: Pricing
    link: https://groq.com/pricing/
  - name: Documentation
    link: https://console.groq.com/docs/overview
  - name: Console
    link: https://console.groq.com/keys
models:
  - name: Gemma 2
    id: gemma2-9b-it
    description: Lightweight, open-weight model well-suited for a variety of text generation tasks. From Google, and built from the same research and technology used to create the Gemini models.
    capabilities:
      tool: true
    input:
      text: true
    output:
      text: true
    context:
      - window: 8_192
  - name: Llama 3.3 Versatile
    id: llama-3.3-70b-versatile
    description: Meta's most advanced multilingual large language model, optimized for a wide range of natural language processing tasks. High performance and efficiency.
    capabilities:
      tool: true
      thinking: true
    input:
      text: true
    output:
      text: true
    context:
      - window: 128_000
      - output: 32_768
  - name: Llama 3.1 8B Instant
    id: llama-3.1-8b-instant
    description: Model with a balance of speed and performance with significant cost savings compared to larger models. Low-latency, high-quality responses.
    capabilities:
      tool: true
    input:
      text: true
    output:
      text: true
    context:
      - window: 128_000
      - output: 8_192
  - name: Llama 3 70b 8192
    id: llama3-70b-8192
    description:  Model with a balance of performance and speed as a reliable foundation model that excels at dialogue and content-generation for tasks requiring smaller context windows.
    capabilities:
      tool: true
    input:
      text: true
    output:
      text: true
    context:
      - window: 8_192
  - name: Llama 3 8B 8192
    id: llama3-8b-8192
    description: Economical model with great performance, industry-leading speed, and cost-efficiency.
    capabilities:
      tool: true
    input:
      text: true
    output:
      text: true
    context:
      - window: 8_192