FutureAItechXFutureAITechX
Explore AI Tools
Groq · Inference API

GroqCloud

Low-latency inference API for open models running on Groq's LPU hardware.

Inference & Model HostingLLM APIs & Model Platforms
Provider
Groq
Latest confirmed model
Not applicable / not verified
Pricing
Free plan + paid
Free plan
Yes
API
Yes · free quota
Status
Generally available
Last verified
Oct 19, 2026
Visit official website Pricing page API docs

What is GroqCloud?

GroqCloud is an API for running open models, such as OpenAI's gpt-oss and Meta's Llama, on Groq's custom LPU chips designed for fast token generation. It offers a free plan with rate limits and a Developer plan with higher limits plus batch and flex processing.

Getting fast responses from open models for real-time apps without hosting them.

Why use GroqCloud?

Practical benefits

  • Focus on fast inference
  • Free plan to start
  • Low per-token prices for gpt-oss models
  • Speech-to-text with Whisper

Who it suits

  • Developers building real-time chat or voice apps
  • Prototypers wanting a free API key

When another tool may fit better

  • Users who need proprietary frontier models
  • Workloads needing very large context windows (production models list ~131K)
  • Teams needing published prices for every model (some say contact sales)
Learning curve
Beginner-friendly
Skills needed
Basic API usage.

Example workflows

  1. Get a key in the Groq console and call openai/gpt-oss-20b
  2. Transcribe audio with whisper-large-v3-turbo

What can it do?

  • Production modelsIncluding openai/gpt-oss-120b, gpt-oss-20b, Llama 3.1 8B, Llama 3.3 70B and Whisper.
  • Preview models PreviewIncluding Qwen 3.8 27B, MiniMax M2.7 and Orpheus TTS; not for production use.
  • Free plan LimitedRate-limited access; check limits in account settings.
  • Developer plan Depends on planHigher limits and Batch and Flex processing.
  • Speech-to-textWhisper large v3 at $0.111 per hour; v3 turbo at $0.04 per hour.

Good for

  • Real-time chatbots
  • Voice assistants
  • Transcription
  • Prompt safety classification

Where it runs

Web (Groq console)REST API

Pricing and plans

Free plan + paid Free plan with rate limits, then usage-based pricing per 1M tokens on the Developer plan. For example, openai/gpt-oss-120b is $0.15 input / $0.60 output and gpt-oss-20b $0.075 / $0.30; some models are contact-sales.

PlanPriceBillingFree trial / tierUsage limitsMain inclusionsModel accessAPI accessRestrictions
FreeFreeusage-based Free tierRate-limited; exact limits shown in account settings
  • API access to models
 Available modelsYesLower limits than Developer plan
DeveloperPer 1M tokens or per audio hour by modelusage-based —Base limits e.g. gpt-oss-120b 30 RPM, 1K RPD, 8K TPM
  • Higher limits
  • Batch processing
  • Flex processing
 All modelsYes—
EnterpriseContact providerusage-based —Higher limits for select workloads — —Yes—

Prices in USD · Global list price · Taxes may apply · Verified Oct 19, 2026

Worked example (usage-based API pricing): sending 1,000,000 input tokens and receiving 200,000 output tokens at the listed openai/gpt-oss-120b rates ($0.15 and $0.60 per 1M) costs about $0.27 before tax. Real bills depend on the model, caching, batch discounts and any tier you qualify for.

Official pricing page (opens in a new tab)

API and developer access

Official API
Yes
Authentication
API key (Bearer)
Pricing model
per 1M tokens; per hour of audio
Free API quota
Free plan with rate limits (values shown in account settings)
Rate limits
Developer base limits e.g. gpt-oss-120b 30 RPM / 1K RPD / 8K TPM
Context / input limits
131,072 tokens on most production models
SDKs
Not documented
Streaming
Not documented
Batch processing
Yes
Structured output
Not documented
Tool / function calling
Not documented
Commercial use
See the provider terms
Data retention & privacy
See the provider privacy terms

Published API rates

ItemPriceUnit
openai/gpt-oss-120b input$0.15per 1M input tokens
openai/gpt-oss-120b output$0.60per 1M output tokens
openai/gpt-oss-20b input$0.075per 1M input tokens
openai/gpt-oss-20b output$0.30per 1M output tokens
whisper-large-v3-turbo$0.04per hour of audio

A free GroqCloud account is not the same thing as free API access. The API allowance above is the provider's documented developer quota.

Verified Oct 19, 2026

Advantages and disadvantages

Advantages

  • Speed-focused hardware
  • Free entry point
  • Simple per-token prices for gpt-oss models

Disadvantages

  • Several models list pricing as 'Contact Sales'
  • Free-plan limits not published in docs
  • Preview models may be discontinued at short notice

Privacy: Not verified in this research.

Real-world examples

A hobbyist trying fast inference

  1. Sign up for the Groq console
  2. Create an API key
  3. Call openai/gpt-oss-20b within free-plan limits

free plan works

A podcast team transcribing episodes

  1. Upload audio to whisper-large-v3-turbo
  2. Pay $0.04 per audio hour
  3. Edit the transcript

needs API key

A developer adding a safety filter

  1. Send user prompts to llama-prompt-guard-2-22m
  2. Block flagged prompts before calling the main model

needs API key

GroqCloud alternatives

Replicate

Run and deploy open-source and custom AI models through an API, paying per second of hardware or per output.

no free plan · offers an API

GroqCloud vs Replicate

Compare GroqCloud with these side by side

Frequently asked questions

Is Groq free?

GroqCloud has a free plan with rate limits. The exact free limits are shown on the Limits page in your account settings.

How much does gpt-oss-120b cost on Groq?

$0.15 per 1M input tokens and $0.60 per 1M output tokens, per Groq's models page.

What does the Developer plan add?

Higher rate limits plus Batch and Flex processing.

Can I use preview models in production?

Groq advises against it, because preview models may be discontinued at short notice.

Does Groq offer transcription?

Yes, Whisper large v3 ($0.111 per hour) and Whisper large v3 turbo ($0.04 per hour).