Groq is an AI solutions company delivering ultra-low latency inference with the first ever LPU™ Inference Engine. Groq API enables developers to integrate state-of-the-art LLMs such as Llama-2 into low latency applications. Learn more at groq.com.
Note: For the following to work with --jsonand jq, the minimum curl version required is v7.82.0. See here for usage with older curl versions.
List available models:
curl -s -H"Authorization: Bearer ${APIKEY}" https://api.groq.com/v1/model_manager/models | jq
Select desired model and generate inference by replacing "model_id" parameter with desired model:
curl -s -H"Authorization: Bearer ${APIKEY}" --json '{"model_id": "llama2-70b-4096", "system_prompt": "You are an unhelpful assistant", "user_prompt": "Are you a fish?"}' https://api.groq.com/v1/request_manager/text_completion | jq
Note: Mixtral is only available to select users at this time. If you are interested, please email us at api@groq.com.