Groq Llama 3.3 70B — LPU hardware, low TTFT champion
Free tier: 30 RPM / 6K TPM. Dev tier: 1000 RPM, 25% discount on tokens. Enterprise: custom. CollabLists 2026-05-20: signed up free tier; tried to upgrade to Dev tier but Groq said no availability. Key works on free tier as 3rd-tier fallback only — 6K TPM too low for primary. Speed: 476 tok/s on gpt-oss-120b, sub-300ms TTFT.
console.groq.com- Author
- Jacob Cole
- Status
- —
- Visibility
- (inherits public)
- Created
- 5/20/2026, 6:47:57 AM
- Updated
- 5/20/2026, 6:47:57 AM
- Permalink
/list/llm-inference-providers/item/366241b9-4165-4634-bee2-5eda7d1b695b