Groq API Pricing and Model Catalog: Use-Case Guide

Groq offers ultra-low latency inference on open-weight models. Here's how to use their API cost-effectively.

|

Pricing data sourced from our catalog. Check data sources for provenance and freshness.

What is Groq?

Groq is an inference provider specializing in:

  • Low latency: Time to first token depends on the model, input length, and service tier. See Groq's latency guide.
  • Open-weight models: Llama, Mixtral, Gemma, and more
  • Speed-optimized: Built for real-time applications
  • Simple pricing: Predictable per-token costs

Groq model catalog

Chat models

Model Input Output Context
llama-3.1-8b-instant
$0.0500 $0.0800 8K
gemma-7b-it
$0.0500 $0.0800 8K
gpt-oss-20b
$0.0750 $0.3000 33K
gpt-oss-safeguard-20b
$0.0750 $0.3000 66K
llama-guard-4-12b
$0.2000 $0.2000 8K

Cost estimation example

Let's estimate costs for a typical Groq chat workload:

Input tokens per request: 500
Output tokens per request: 200
Requests per month: 100,000
Monthly cost: $4.10

Groq strengths

  • Latency-aware selection: Measure time to first token and output speed with your own prompt sizes.
  • Open-weight models: Access to the latest open-weight models
  • Simple pricing: Predictable per-token costs
  • No vendor lock-in: Use the same models elsewhere
  • Real-time applications: Ideal for chatbots and interactive apps

When to use Groq

  • Real-time chat: When latency matters most
  • Interactive applications: When users expect instant responses
  • Open-weight models: When you want model flexibility
  • Cost-sensitive workloads: When you need predictable costs

Cost optimization tips

  • Use smaller models: For less complex tasks
  • Batch API: For workloads that can wait, eligible models cost 50% less than synchronous requests. Batch and prompt-caching discounts do not stack.
  • Caching: Cache responses for repeated queries
  • Right-size your model: Match model size to task complexity
  • Monitor usage: Track token usage to optimize costs

Compare Groq models

Ready to compare Groq models side by side? Use our tools:

Related guides

Frequently asked questions

What is Groq's cheapest model?

Groq's Llama 3.1 8B and Gemma 2 9B models are typically the most cost-effective.

How does Groq compare to direct providers?

Groq offers lower latency but may have higher per-token costs than direct providers like OpenAI or Anthropic.

Does Groq offer batch pricing?

Yes. Groq offers a Batch API with 50% lower pricing than synchronous requests for supported chat and audio models. Jobs run asynchronously with a 24-hour to 7-day processing window; batch and prompt-caching discounts cannot be combined. Check the official Batch API documentation for eligible models and terms.

Pricing data sourced from official provider documentation. Prices may vary by region and usage tier.