Groq is a paid tool. Please visit their website for pricing details.

Groq

Fast AI inference through GroqCloud, with model-specific usage pricing and LPU-based infrastructure.

paid

Last Updated 8/28/2026

About Groq

Groq is a cloud AI inference service built on the company’s LPU architecture. Developers use GroqCloud to run supported openly available models while Groq operates the underlying inference infrastructure. The pricing catalog shows throughput together with separate input and output token rates, so teams can compare expected responsiveness and API cost before choosing a model. Listed model families include GPT OSS, Llama, and Qwen, while some models are limited to enterprise access. Text-to-speech usage is billed by character volume. Community resources and chat in the GroqCloud Developer Console are available for developer support, and enterprise teams can contact Groq about specific inference workloads. Security and compliance documentation is provided through the Groq Trust Center. Vulnerability reports are handled through Groq’s private HackerOne program.

Best for

Developers building latency-sensitive AI applications

Teams comparing model-serving cost by token volume

Enterprises that need to discuss dedicated inference access

Best use cases

Serving interactive assistants with fast token generation

Selecting a supported model against a defined API budget

Evaluating enterprise-only model access with Groq

Compare before choosing

Compare separate input and output rates before choosing a model

Check published throughput because speed differs by model

Confirm enterprise access requirements for restricted models

Features

Fast LLM inference

Run supported openly available language models on Groq’s LPU-based cloud infrastructure with published throughput figures.

Usage-based API pricing

Pay separately for input and output tokens, with model-specific rates published per million tokens.

Model choice

Choose among models including GPT OSS, Llama, and Qwen variants for different speed, context, and cost requirements.

Pros

Published per-model token rates make API cost comparisons straightforward.
High published token throughput suits response-sensitive application workflows.
Multiple supported model families let teams trade off speed, context, and cost.

Cons

Input and output charges differ by model, so cost estimates require workload-specific token volumes.
Some models are enterprise-only and require contacting Groq before adoption.
Model availability and compatibility are limited to the catalog supported by GroqCloud.

Pricing Plans

On-Demand

Model-dependent usage ratesper million tokens or modality unit

GPT OSS 20B: $0.075 input and $0.30 output per 1M tokens
Llama 3.1 8B Instant: $0.05 input and $0.08 output per 1M tokens
Qwen 3.6 27B: $0.60 input and $3.00 output per 1M tokens

Enterprise-only models

Contact salescustom

Enterprise access inquiry
Minimax M2.7 listed as enterprise-only
Pricing arranged with Groq

* Pricing information is for reference only. Please refer to the latest announcement on Groq's official website.

FAQ

Groq is an AI inference provider. GroqCloud uses the company’s LPU architecture to serve supported models with an emphasis on speed and predictable usage-based cost.

On-demand models are priced by usage. Language-model input and output tokens have separate per-million-token rates, while other modalities can use units such as characters.

Developers can use Groq community resources or the chat feature in the GroqCloud Developer Console. Enterprise buyers can contact the company about their inference use case.