HTTP 200 verified twice
AI infrastructure · Tool
Groq
ResearchedGroq delivers fast, low-cost AI inference through its custom LPU silicon, pioneered in 2016 as the first chip purpose-built for inference. Its GroqCloud platform serves 3M developers and enterprises via tokens-as-a-service pricing, OpenAI-compatible APIs, and global data center deployment.
Official site snapshots
Visit live site See the official site at a glance
Read-only public captures of Groq’s homepage. Screenshots are dated, never live embeds, and open full-screen.
In one minute
Start here for the decision-making essentials: what Groq does, who it is for, how it is accessed, and the first-party sources behind this profile.
Pricing4 options
Tokens-as-a-Service on-demand pricing with per-million-token input and output rates
GPT OSS 120B (128k context) priced at $0.15 input / $0.60 output per million tokens at 500
Llama 3.1 8B Instant (128k context) priced at $0.05 input / $0.08 output per million token
Llama 3.3 70B Versatile (128k context) priced at $0.59 input / $0.79 output per million to
Platforms
GroqCloud powers 1M+ developers
API accessNot public
FoundedGroq pioneered the LPU in 2016 as the first chip purpose-built for inference.
AvailabilityWeb / remote
LicenseProprietary
Best suited to
Source-backed fit Developers building latency-sensitive AI applications Enterprises needing large-scale inference capacity and dedicated support Teams migrating from OpenAI with minimal code changes Startups running inference-based AI businesses rather than training LLMs High-volume workloads suited to async batch processing Serves both developers and enterprises
Decision support
6 sourced answersCommon questions and adoption checks
Short answers to the questions buyers and builders commonly ask about Groq. Each answer cites the shared ledger below, where every source is listed once.
01What does Groq say it can do?
Groq delivers fast, low-cost inference via the LPU. · Fast, low-cost inference delivered by the Groq LPU · Fast, low cost inference via the Groq LPU for developers · Fast, low-cost AI inference
Groq is fast, low cost inference. The Groq LPU delivers inference with the speed and cost developers need.
02Who is Groq intended for?
Serves both developers and enterprises · Fortune 500 companies and more than 1.4 million developers building real-time AI applicati
fast, low latency API access to state-of-the-art models for everyone from the developer to the enterprise
03What use cases does Groq describe?
Running inference-based AI businesses (as opposed to training LLMs) · AI-generated text detection · Industry solutions for Healthcare, Defense, Finance, Entertainment, Gaming, and Telecommun · Retrieval Augmented Generation (RAG)
For training LLMs, there are options. But to run an inference-based AI business, you need something different. Groq was the right choice.
04What should teams verify before adopting Groq?
Cached tokens do not count towards GroqCloud rate limits
All cached tokens don't count towards GroqCloud rate limits.
05What pricing information is available for Groq?
Tokens-as-a-Service on-demand pricing with per-million-token input and output rates · GPT OSS 120B (128k context) priced at $0.15 input / $0.60 output per million tokens at 500 · Llama 3.1 8B Instant (128k context) priced at $0.05 input / $0.08 output per million token · Llama 3.3 70B Versatile (128k context) priced at $0.59 input / $0.79 output per million to
Groq On-Demand Pricing for Tokens-as-a-Service | Groq is fast, low cost inference.
06Does Groq document API access?
A free API key is available to developers · Developer Tier offers rate limits up to 10x higher than the free tier. · Free API key available for developers · OpenAI-compatible API endpoint at https://api.groq.com/openai/v1/responses, allowing migra
Free API key
Decision guide
AI infrastructureCapabilities and operating fit
This profile connects the jobs Groq is described as handling with its delivery model, access options and the subjects used to match it to related products in this directory.
Common use cases
- Running inference-based AI products at low latency worldwide
- Pay-as-you-go inference via Developer Tier with a credit card
- Migrating from OpenAI by changing three lines of code
- High-volume parallel API workloads via the Batch API
- Tool-enabled AI applications through Remote MCP integrations
- Web automation, search, scraping, and invoicing via MCP servers like BrowserBase, Exa, Fir
Topics mapped
Verified capabilities
- Groq delivers fast, low-cost inference via the LPU.
- Fast, low-cost inference delivered by the Groq LPU
- Fast, low cost inference via the Groq LPU for developers
- Fast, low-cost AI inference
- Fast, low-cost inference
- Groq LPU delivers inference with the speed and cost developers need
- Fast, low cost inference at scale via the Groq LPU
- 99.5% uptime / platform availability
Recorded integrations
GroqCloud offers OpenAI endpoint compatibility, allowing migration by changing three linesRemote Model Context Protocol (MCP) server integration available in Beta on GroqCloud, conSupported MCP server integrations include BrowserBase, Browser Use, Exa, Firecrawl, HugginRemote MCP connectors/support are available in beta on GroqCloudGroq integrates with Hugging Face for building fasterOfficial Llama API accelerated by Groq in collaboration with MetaGroq API can connect proprietary data to open-source LLMsWorks with vector databases such as Pinecone for unstructured data retrievalMigrating from OpenAI to Groq requires only three lines of code
Intended audiences
Access signals
- Pricing model
- Tokens-as-a-Service on-demand pricing with per-million-token input and output rates · GPT OSS 120B (128k context) priced at $0.15 input / $0.60 output per million tokens at 500 · Llama 3.1 8B Instant (128k context) priced at $0.05 input / $0.08 output per million token · Llama 3.3 70B Versatile (128k context) priced at $0.59 input / $0.79 output per million to
- API
- Not publicly listed
- Source links
- 17 recorded
Source-backed
Updated July 22, 2026Verified facts
Each fact points to a recorded source, making it easy to distinguish verified product information from claims that need checking.
Groq is fast, low cost inference.
Fast, low-cost inference delivered by the Groq LPU · Groq pioneered the LPU in 2016, the first chip purpose-built for inference · Custom silicon rather than GPUs alone
Groq delivers fast, low-cost inference via the LPU.
Groq pioneered the LPU in 2016 as the first chip purpose-built for inference.
Groq is used by 3M developers and teams.
GroqCloud is the cloud platform/console for running inference on Groq's LPU-based stack.
Groq uses its own custom LPU silicon rather than relying on GPUs alone.
View 32 more verified facts
Groq's LPU-based stack runs in data centers across the world for low-latency inference.
Tokens-as-a-Service on-demand pricing with per-million-token input and output rates
A free API key is available to developers
GPT OSS 120B (128k context) priced at $0.15 input / $0.60 output per million tokens at 500 TPS
Llama 3.1 8B Instant (128k context) priced at $0.05 input / $0.08 output per million tokens at 840 TPS
Llama 3.3 70B Versatile (128k context) priced at $0.59 input / $0.79 output per million tokens at 394 TPS
Developer Tier is a self-serve, pay-as-you-go plan requiring only a credit card to access GroqCloud.
Developer Tier offers rate limits up to 10x higher than the free tier.
The Batch API provides a 25% cost discount with guaranteed 24-hour processing time for high-volume parallel workloads.
Flex Tier (beta) supports llama-3.3-70b-versatile and llama-3.1-8b-instant models with up to 10x rate limits.
GroqCloud offers OpenAI endpoint compatibility, allowing migration by changing three lines of code.
Groq provides an Enterprise tier for large-scale needs with enterprise-grade capacity and dedicated support.
Fast, low-cost inference delivered by the Groq LPU
2016
Groq is focused on inference
Free API key available for developers
Developer support via community resources and GroqCloud Dev Console chat; enterprise inquiries handled by sales team
Fast, low cost inference via the Groq LPU for developers
Groq is committed to protecting customers, systems, and data
Proactively identifies and mitigates security risks using industry-standard practices
Operates a private vulnerability disclosure program through HackerOne
Eligible vulnerability submissions may be recognized or rewarded at Groq's discretion
Groq Trust Center provides information on the security program, compliance posture, and supporting documentation
Groq was established in 2016 focused on inference
Fast, low-cost AI inference
Serves both developers and enterprises
Sep 23, 2025
Remote Model Context Protocol (MCP) server integration available in Beta on GroqCloud, connecting AI models to thousands of external tools through Anthropic's open MCP standard
OpenAI-compatible API endpoint at https://api.groq.com/openai/v1/responses, allowing migration from OpenAI with zero code changes
Remote MCP enables developers to connect any remote MCP server to models hosted on GroqCloud, enabling tool capabilities
Remote MCP is available on all models that support tool use, including openai/gpt-oss-20b, openai/gpt-oss-120b, moonshotai/kimi-k2-instruct-0905, qwen/qwen3-32b, meta-llama/llama-4-maverick-17b-128e-instruct, meta-llama/llama-4-scout-17b-16
Supported MCP server integrations include BrowserBase, Browser Use, Exa, Firecrawl, HuggingFace, Parallel, Stripe, and Tavily
Practical capabilities
8 documented areasWhat it helps with
A concise view of the jobs, capabilities and integrations described in the recorded product sources.
Running inference-based AI products at low latency worldwide
Pay-as-you-go inference via Developer Tier with a credit card
Migrating from OpenAI by changing three lines of code
High-volume parallel API workloads via the Batch API
Tool-enabled AI applications through Remote MCP integrations
Web automation, search, scraping, and invoicing via MCP servers like BrowserBase, Exa, Fir
Enterprise-grade custom inference solutions
Running inference-based AI businesses (as opposed to training LLMs)
Availability
Source checkedWhere it runs and where to get it
Documented product formats, platforms and official distribution destinations. Availability can vary by region and plan.
Cost / license
Tokens-as-a-Service on-demand pricing with per-million-token input and output rates · GPT OSS 120B (128k context) priced at $0.15 input / $0.60 output per million tokens at 500 · Llama 3.1 8B Instant (128k context) priced at $0.05 input / $0.08 output per million token · Llama 3.3 70B Versatile (128k context) priced at $0.59 input / $0.79 output per million toGroq API Cookbook is an open-source community repository · compound-beta integrates multiple open-source models
Platforms
GroqCloud is the cloud platform/console for running inference on Groq's LPU-based stack.GroqCloud powers 1M+ developersGroq LPU, a custom-built inference chip, delivered via GroqCloudGroqCloud™ Dev Console is the platform where models from the Llama 3.2 suite are made avai
Adoption notes
DeploymentGroq's LPU-based stack runs in data centers across the world for low-latency inference.
LicenseGroq API Cookbook is an open-source community repository · compound-beta integrates multiple open-source models
Model supportFlex Tier (beta) supports llama-3.3-70b-versatile and llama-3.1-8b-instant models with up · Remote MCP is available on all models that support tool use, including openai/gpt-oss-20b, · Llama 4 inference is available live on GroqCloud · Llama 4 (Meta) available via the official Llama API on Groq · compound-beta and compound-beta-mini
Data controlGroq is committed to protecting customers, systems, and data · Proactively identifies and mitigates security risks using industry-standard practices · Operates a private vulnerability disclosure program through HackerOne · Eligible vulnerability submissions may be recognized or rewarded at Groq's discretion · Compliance-ready inference for handling classified defense documents, protected health inf
Learning curveIntermediate
Primary use casesRunning inference-based AI products at low latency worldwide, Pay-as-you-go inference via Developer Tier with a credit card, Migrating from OpenAI by changing three lines of code, High-volume parallel API workloads via the Batch API, Tool-enabled AI applications through Remote MCP integrations, Web automation, search, scraping, and invoicing via MCP servers like BrowserBase, Exa, Fir, Enterprise-grade custom inference solutions, Running inference-based AI businesses (as opposed to training LLMs), AI-generated text detection, Industry solutions for Healthcare, Defense, Finance, Entertainment, Gaming, and Telecommun, Retrieval Augmented Generation (RAG), Building real-time AI agents and agentic workflows with minimal configuration, Production AI workloads and real-time AI applications requiring consistent low latency and, The Llama 3.2 11B and 90B vision models on Groq support image reasoning use cases.
What to verify before adopting
- Cached tokens do not count towards GroqCloud rate limits
Evolution and major updates
Research in progressGroq timeline
A concise history of software releases and material product changes. Events appear only when a dated source supports what changed.
Scheduled for research
Building a reliable release history.
This profile is being checked for dated releases and material product changes. Nothing appears here until the exact date and event can be verified from a recorded source.
Dated event Short explanation Original source
Citation ledger
17 unique pagesRecorded sources
Facts, answers, structured details, milestones and primary resource links cite this shared ledger. Each external page appears once; release tags from the same GitHub project are grouped under one release history.
- 1groq.com 10 facts · 1 answer · Official site
- 2groq.com/blog 8 facts · 1 answer · Official site
- 3groq.com/blog/developer-tier-now-available-on-groqcloud 6 facts · 2 answers · Official site
- 4groq.com/blog/gpt-oss-improvements-prompt-caching-and-lower-pricing 6 facts · 2 answers · Pricing
- 5groq.com/blog/how-to-build-your-own-ai-research-agent-with-one-groq-api-call 6 facts · 1 answer · Documentation
- 6groq.com/blog/retrieval-augmented-generation-with-groq-api 6 facts · 1 answer · Documentation
- 7groq.com/changelog 6 facts · 1 answer · Official site
- 8groq.com/contact 5 facts · 2 answers · Official site
- 9groq.com/customer-stories 6 facts · 1 answer · Official site
- 10groq.com/industry-solutions 6 facts · 1 answer · Official site
- 11groq.com/pricing 5 facts · 2 answers · Pricing
- 12groq.com/security 6 facts · 1 answer · Security
- 13groq.com/blog/meta-and-groq-continue-to-build-open-source-developer-ecosystem-as-llama-3-2-launches 6 facts · Official site
- 14groq.com/newsroom/meta-and-groq-collaborate-to-deliver-fast-inference-for-the-official-llama-api 5 facts · 1 answer · Documentation
- 15groq.com/blog/groq-vercel-partner-to-make-building-fast-and-simple 3 facts · 2 answers · Official site
- 16groq.com/blog/thank-you-1m-developers-building-with-groqcloud 5 facts · Documentation
- 17groq.com/blog/the-official-llama-api-accelerated-by-groq 5 facts · Documentation
Research status99 substantive facts · 17 source pages · quality score 100/100
Connected directory
Semantic matchesExplore related solutions
These tools share a workflow, capability or audience with Groq. They may complement it rather than replace it, and are not presented as integrations or endorsements.
