Groq
API
30 requests/min, 14,400 requests/day, 15,000 tokens/min, 500,000 tokens/day.
Reset · Provider minute, hour (audio where applicable), and day windows. See the account Limits page for actual limits and the response headers for retry/reset timing.
Setup, evidence & eligibility
Limits are enforced at the organization level; Groq directs users to account settings for exact limits. Groq lists this as a preview for evaluation.
How to use it
- Sign in to GroqCloud, create a key under API Keys, and check your organization’s Free Plan limits.
- Use the exact model ID listed below.
- Create a Groq API key on the Free plan.
- Send messages and model meta-llama/llama-prompt-guard-2-22m to the chat completions endpoint.
- Check your current organization limits in GroqCloud.
- Model ID
meta-llama/llama-prompt-guard-2-22m- Endpoint
https://api.groq.com/openai/v1/chat/completions- Protocol
OpenAI-compatible REST with a Groq API key
Checks & requirements
- Listing status
- Documented
- Source checked
- Nov 18, 2026
- Hands-on tested
- Not recorded
- Account checked
- Not recorded
- Generation tested
- Not recorded
- Account
- Required
- Payment card
- Unknown
- Phone
- Unknown
- Regions
- Free-plan availability in India has not been verified.
- India access
- Unknown
- Watermark
- Not applicable
- Commercial use
- Unknown
- Privacy
- Groq generally does not retain standard inference inputs or outputs. Reliability and abuse review can trigger temporary logging. See Data Controls.
- Groq Free Plan limits Provider documentation · Checked Nov 18, 2026
The named Free Plan row lists 30 requests/min, 14,400 requests/day, 15,000 tokens/min, 500,000 tokens/day, subject to organization-level exceptions.
- Groq model and usage docs Provider documentation · Checked Nov 18, 2026
Classify input for prompt injection with the smaller Prompt Guard checkpoint.
- Groq API quickstart Provider documentation · Checked Nov 18, 2026
Create a GroqCloud API key for authenticated API requests.
- Groq data controls Provider documentation · Checked Nov 18, 2026
Default inference does not retain inputs or outputs, with exceptions for reliability, abuse monitoring and stateful features.
- Free Inference source snapshot Directory report · Checked Nov 18, 2026
Groq was identified by Free Inference. Exact availability and quota come from Groq's current Free Plan table.
- Groq Cloud platform Provider documentation · Checked Nov 18, 2026
Groq's product page links to its official hosted inference platform.