About Groq
Groq runs AI model inference on its own LPU hardware, with a focus on fast responses. Developers reach the models through an API hosted in the Groq cloud.
What it does
Groq's offer centers on speed and cost at scale. Developers send requests to supported models through an API and get responses back quickly. The company also describes work with other hardware partners, so its capacity can grow as demand increases.
Key features
- Fast inference on LPU hardware
- API access to hosted models
- Capacity built for large workloads
- Developer console for keys and usage
Who it's for
- Developers building apps where response time matters
- Teams running agents that call a model many times in a row
- Companies planning high-volume inference
Pricing and good to know
Pricing is set per model and per usage, so compare the rates for the model you plan to use. Free access may be available for testing, but limits and availability can change. Check the console for current quotas before you plan production traffic, since demand on fast hardware can affect limits.
Highlights
- Fast inference on LPU hardware
- Call hosted models through an API
- Run large workloads at scale
- Manage keys and usage in a console
