Groq

Groq is a high-speed AI inference platform that runs large language models at exceptionally fast response rates — often 10-20x faster than comparable cloud services. Built on custom Language Processing Units (LPUs), Groq hosts models including Llama 3.1 8B, Llama 3.3 70B, Llama 4 Scout, OpenAI's GPT-OSS 20B/120B, Alibaba's Qwen3 32B, Moonshot AI's Kimi K2, and Whisper Large v3 for speech, accessible via a chat playground or API. It is popular among developers who need fast prototyping, low-latency applications, or simply want to experience what frictionless LLM inference feels like. The free tier requires no credit card and is governed by generous per-minute rate limits rather than a monthly token budget; paid API access is billed per token (as low as $0.05/1M input tokens for Llama 3.1 8B), with a 50% Batch API discount and automatic prompt-caching discounts.