On-demand Pricing for Tokens-as-a-Service

Groq powers leading openly-available AI models.

Get started for free and upgrade as your needs grow. View the pricing of our core models below – note all prices are in USD. Other models are available for specific customer requests including fine tuned models. Send us your inquiries here.

Large Language Models (LLMs)

*Approximate number of tokens per $
AI Model	Current Speed(Tokens per Second)	Input Token Price(Per Million Tokens)	Output Token Price(Per Million Tokens)
GPT OSS 20B 128k	1,000	$0.10(10M / $1)*	$0.50(2M / $1)*	Try Now Model Card
GPT OSS 120B 128k	500	$0.15(6.67M / $1)*	$0.75(1.33M / $1)*	Try Now Model Card
Kimi K2 1T 128k	200	$1.00(1M / $1)*	$3.00(333,333 / $1)*	Try Now Model Card
Llama 4 Scout (17Bx16E) 128k	594	$0.11(9.09M / $1)*	$0.34(2.94M / $1)*	Try Now Model Card
Llama 4 Maverick (17Bx128E) 128k	562	$0.20(5M / $1)*	$0.60(1.6M / $1)*	Try Now Model Card
Llama Guard 4 12B 128k	325	$0.20(5M / $1)*	$0.20(5M / $1)*	Try Now Model Card
DeepSeek R1 Distill Llama 70B 128k	400	$0.75(1.33M / $1)*	$0.99(1.01M / $1)*	Try Now Model Card
Qwen3 32B 131k	662	$0.29(3.44M / $1)*	$0.59(1.69M / $1)*	Try Now Model Card
Mistral Saba 24B 32k	330	$0.79(1.27M / $1)*	$0.79(1.27M / $1)*	Try Now
Llama 3.3 70B Versatile 128k	394	$0.59(1.69M / $1)*	$0.79(1.27M / $1)*	Try Now Model Card
Llama 3.1 8B Instant 128k	840	$0.05(20M / $1)*	$0.08(12.5M / $1)*	Try Now Model Card
Llama 3 70B 8k	330	$0.59(1.69M / $1)*	$0.79(1.27M / $1)*	Try Now Model Card
Llama 3 8B 8k	1,345	$0.05(20M / $1)*	$0.08(12.5M / $1)*	Try Now Model Card
Gemma 2 9B 8k	500	$0.20(5M / $1)*	$0.20(5M / $1)*	Try Now Model Card
Llama Guard 3 8B 8k	765	$0.20(5M / $1)*	$0.20(5M / $1)*	Try Now Model Card

Text-to-Speech (TTS) Models

AI Model	Characters /s	PricePrice (Per M Characters)
PlayAI Dialog v1.0	140	$50.00	Try Now Model Card

Automatic Speech Recognition (ASR) Models

*Audio is billed at a minimum of 10s per request.
AI Model	Speed Factor	Price(Per Hour Transcribed)
Whisper V3 Large	217x	$0.111*	Try Now Model Card
Whisper Large v3 Turbo	228x	$0.04*	Try Now Model Card

Built In Tools (Compound)

Tool	Price	Parameter
Basic Search	$5 / 1000 requests	web_search
Advanced Search	$8 / 1000 requests	web_search
Visit Website	$1 / 1000 requests	visit_website
Code Execution	$0.18 / hour	code_interpreter

Built In Tools (GPT-OSS)

Tool	Price	Parameter
Browser Search - Basic Search	$5 / 1000 requests	browser_search - browser.search
Browser Search - Visit Website	$1 / 1000 requests	browser_search - browser.open
Code Execution - Python	$0.18 / hour	code_interpreter - python

Prompt Caching

Note: No extra fee for the caching feature itself. The discount only applies when a cache hit occurs.
Model	Uncached Input Tokens (Per Million Tokens)	Cached Input Tokens (Per Million Tokens)	Output Tokens (Per Million Tokens)
moonshotai/kimi‑k2‑instruct	$1.00	$0.50	$3.00

Batch API

Batch processing lets you run thousands of API requests at scale by submitting your workload as an asynchronous batch of requests to Groq with 50% lower cost, no impact to your standard rate limits, and 24-hour to 7 day processing window.

Learn more about Batch pricing and how to get started.

For enterprise API solutions or on-prem deployments, please fill out the form on our Enterprise Access Page.

Compound Systems

Compound Systems

Compound AI systems are powered by multiple openly-available models already supported in GroqCloud to intelligently and selectively use tools to answer user queries, starting first with web search and code execution.Pricing is passed through to the underlying models and server side tools that are part of the compound AI system. While in beta, tool calls for Compound AI Systems are not charged.

For more information, see the GroqCloud documentation.