What is GroqCloud
GroqCloud is a hosted inference platform built around Groq's language processing units, offering APIs for text, speech and compound tool-using systems.
Plans and pricing
As of September 7, 2026, the Free tier provides limited no-cost access, while Developer is pay as you go with no upfront plan fee. Current production text-model rates per million tokens are GPT-OSS 20B at US$0.075 input and US$0.30 output and GPT-OSS 120B at US$0.15 input and US$0.60 output. Batch processing is priced 50% below standard rates.
Whisper Large V3 costs US$0.111 per audio hour and Whisper Large V3 Turbo costs US$0.04 per hour. Orpheus English text to speech costs US$22 per million characters. Compound tools are billed separately, including basic web search at US$5 per thousand calls and advanced search at US$8 per thousand calls.
Enterprise options
Enterprise and on-premises arrangements use custom pricing and can add higher limits, support and deployment controls. Developer accounts are billed for measured usage, with monthly or progressive charging described in the billing documentation.
Official sources: Groq pricing and billing FAQ.