# Grok vs Groq: xAI's chatbot and Groq's inference cloud explained

> Grok is xAI's assistant and model family, with plans from free to $200 a month. Groq sells fast hosting of other labs' open models, and it does not serve Grok.

URL: https://claudegpt.si/news/grok-vs-groq  
Published: 2026-11-16T03:07:44.285Z  
Author: ClaudeGPT Si Newsroom  
Section: Reviews  
Topics: Grok, xAI, Groq, Comparison, AI chips, Inference platforms  
Publisher: ClaudeGPT Si

## Key takeaways

- Grok is xAI's AI assistant and model family; Groq is a separate company that makes LPU inference chips and runs the GroqCloud API.
- As of November 16, 2026, Grok plans run from free to $200 a month, and the xAI API charges $2 per million input tokens for grok-4.7.
- GroqCloud serves open models such as gpt-oss-120b at $0.15 per million input tokens, and its model list includes no Grok model.

Grok vs Groq is not a contest between rivals: they are two different companies' products that happen to sound alike. Grok is the AI assistant and model family from xAI, which now signs grok.com and its developer docs as SpaceXSI. Groq is an unrelated chip and cloud company: it designed its own inference chip, the LPU, and runs GroqCloud, an API that serves other companies' open models at high speed. You chat with Grok. You build on Groq.

The mix-up matters most to developers. If you want Grok's models in your app, you need an xAI API key. A Groq key will not get you Grok, because as of November 16, 2026 GroqCloud does not serve any Grok model.

## Grok vs Groq at a glance

This comparison is based on the companies' own docs, pricing pages and press releases. All prices are as of November 16, 2026.

| | Grok | Groq |
| --- | --- | --- |
| Made by | xAI (signs as SpaceXSI) | Groq LLC, a private company in San Francisco |
| What it is | A chatbot and a family of AI models | Inference hardware (the LPU) and a cloud API, GroqCloud |
| Whose models | xAI's own Grok models | Other labs' open models: OpenAI's gpt-oss, Alibaba's Qwen, Meta's Llama and more |
| Where you use it | grok.com, the Grok apps, X, Grok Bot, the xAI API | GroqCloud console and API at api.groq.com |
| Consumer app | Yes | No |
| Free option | Free plan with limited usage | Free API plan with rate limits |
| Paid entry price | Plus at $8 a month | Pay per token, from $0.075 per million input tokens (gpt-oss-20b) |
| Flagship API price | grok-4.7: $2 input, $6 output per million tokens | gpt-oss-120b: $0.15 input, $0.60 output per million tokens |
| Best for | People who want an assistant; developers who want Grok models | Developers who want fast, cheap inference on open models |

## What Grok is and who makes it

Grok is xAI's AI assistant. Its [home page](https://grok.com/) describes it as an assistant that can chat, create images, write code and pull real-time answers from the web and X. You can use it on grok.com, in the Grok apps, inside X, and through Grok Bot, xAI's agent that works on your computer and apps.

The same name covers the models behind the assistant. xAI sells them to developers through its own API at api.x.ai. Its [models and pricing docs](https://docs.x.ai/developers/pricing) list grok-4.7, grok-4.6, grok-4.5, grok-4.3, the grok-4.20 family and grok-build-0.1, a coding model.

xAI's documentation now calls the company "SpaceXSI (xAI)", and grok.com says Grok is "built by SpaceXSI". The product name, Grok, has not changed.

## What Groq is and who makes it

Groq is a hardware and cloud company. Its main product for developers is GroqCloud, an API that runs other companies' models. Groq's [docs](https://console.groq.com/docs/overview) describe it as "Fast LLM inference, OpenAI-compatible", which means code written for OpenAI's API can usually point at Groq by changing the base URL and key.

Inference is the work of running a trained model to answer requests, as opposed to training it. Groq built a chip for that job, the LPU. Its [platform page](https://groq.com/platform) sells three layers on top of its hardware: GroqMetal (dedicated bare-metal machines), GroqCore (a managed inference stack) and GroqAssured (enterprise governance and audit controls).

Groq does not offer a chatbot for the public. Its pitch is speed and price. The [GroqCloud models page](https://console.groq.com/docs/models) lists gpt-oss-20b at about 1,000 tokens per second and gpt-oss-120b at about 500.

## Who owns Groq and what changed in 2026

Groq has had a busy year, and its own newsroom tells the story:

- **December 24, 2025:** Groq [announced](https://groq.com/newsroom/groq-and-nvidia-enter-non-exclusive-inference-technology-licensing-agreement-to-accelerate-ai-inference-at-global-scale) a non-exclusive licensing agreement with Nvidia for its inference technology. Founder Jonathan Ross, President Sunny Madra and other team members joined Nvidia. Simon Edwards became CEO, and Groq said GroqCloud would keep running without interruption.
- **June 22, 2026:** Groq [raised $650 million](https://groq.com/newsroom/groq-raises-usd650m-to-scale-its-ai-inference-cloud-business) in a round led by Disruptive and Infinitum. The release names Adam Winter as CEO, Matt Eng as CFO and Alex Davis of Disruptive as chairman. It also says Nvidia announced its LPX platform, which incorporates Groq's inference technology, at this year's GTC.
- **August 12, 2026:** Groq [joined the NVIDIA Cloud Partner program](https://groq.com/newsroom/groq-becomes-an-nvidia-cloud-partner), which certifies it to run Nvidia hardware in its data centers.
- **August 17, 2026:** Groq LLC [announced a $350 million Series A](https://groq.com/newsroom/groq-closes-usd350-million-series-a-building-the-world-s-leading-ai-inference-cloud) at a $3.5 billion valuation, led by Disruptive with planned participation from Nvidia. The company said the round was subject to customary closing conditions.

Groq says it runs 13 data centers across North America, Europe, the Middle East and Asia Pacific and serves more than six million developers. It has not published who owns what share of the company.

None of this connects Groq to xAI, beyond one hire: Groq's June release says its new COO, Alan Rice, previously worked at xAI.

## How much Grok costs

Grok's consumer plans now come through The X Subscription, which bundles Grok, Cursor and X Premium benefits. These are xAI's list prices from its [pricing docs](https://docs.x.ai/x-subscription/pricing), as of November 16, 2026:

| Plan | Monthly | Annual | Included usage | Grok Bot |
| --- | --- | --- | --- | --- |
| Free | $0 | $0 | Limited | No |
| Plus | $8 | $80 ($7 a month) | Base | No |
| Premium | $30 | $300 ($25 a month) | 4x Plus | Yes |
| Super | $100 | $1,000 ($84 a month) | 15x Plus | Yes |
| Ultra | $200 | $2,000 ($167 a month) | 50x Plus | Yes |

Grok and Cursor draw from one shared weekly usage pool. Team seats cost the same per seat on Premium, Super and Ultra. Existing SuperGrok and X Premium plans continue unless you switch. For a plan-by-plan breakdown, see [Grok Bot pricing](https://claudegpt.si/news/grok-bot-pricing-plans).

Developers pay per token on the xAI API. Selected text models, per million tokens, for prompts under 200,000 tokens, as of November 16, 2026:

| Model | Context window | Input | Cached input | Output |
| --- | --- | --- | --- | --- |
| grok-4.7 | 500k | $2.00 | $0.50 | $6.00 |
| grok-4.3 | 1M | $1.25 | $0.20 | $2.50 |
| grok-4.20-0309-non-reasoning | 1M | $1.25 | $0.20 | $2.50 |
| grok-build-0.1 | 256k | $1.00 | $0.20 | $2.00 |

Prompts of 200,000 tokens or more are billed at double these rates for every token in the request. Server-side tools cost extra: web search is $5 per 1,000 calls.

## How much Groq costs

GroqCloud has a free plan and a paid Developer plan. The free plan comes with rate limits: on gpt-oss-120b, for example, its [rate limits page](https://console.groq.com/docs/rate-limits) allows 30 requests a minute, 1,000 requests a day and 200,000 tokens a day. The Developer plan raises the limits and adds batch and flex processing.

Paid usage is billed per token. Prices from the [GroqCloud models page](https://console.groq.com/docs/models), as of November 16, 2026:

| Model | Developer | Speed (tokens/sec) | Input per 1M | Output per 1M |
| --- | --- | --- | --- | --- |
| openai/gpt-oss-20b | OpenAI | 1,000 | $0.075 | $0.30 |
| openai/gpt-oss-120b | OpenAI | 500 | $0.15 | $0.60 |
| qwen/qwen3.8-27b (preview) | Alibaba Cloud | 450 | $0.80 | $4.00 |
| llama-3.3-70b-versatile | Meta | 280 | Contact sales | Contact sales |
| llama-3.1-8b-instant | Meta | 560 | Contact sales | Contact sales |

Groq lists cached input for gpt-oss-120b at $0.075 per million tokens. Its two Llama models are marked Enterprise, with no public price. Speech-to-text is priced by the hour: $0.111 for whisper-large-v3 and $0.04 for whisper-large-v3-turbo.

The two price lists do not compare like for like. Grok's prices buy xAI's own frontier models. Groq's buy fast hosting of open models that other labs released.

## Can you run Grok on Groq

No. As of November 16, 2026, GroqCloud's [model list](https://console.groq.com/docs/models) has no Grok model in either its production or preview tier. Its models come from Meta, OpenAI, Alibaba Cloud, MiniMax and Canopy Labs.

To use Grok in code, you need an xAI API key and the api.x.ai endpoint. To use Groq, you need a GroqCloud key and api.groq.com. Each key works only on its own service.

## Which one you want

If you want an assistant to talk to, you want Grok. Start on its free plan, or pay $8 a month for Plus. If you are choosing between chatbots, our [ChatGPT vs Grok comparison](https://claudegpt.si/news/chatgpt-vs-grok) covers that decision.

If you are a developer who needs answers back fast and cheap, and an open model like gpt-oss is good enough, Groq is built for that. If you need Grok's own models specifically, only xAI sells them.

## Frequently asked questions

### Is Groq the same as Grok?

No. Grok is the AI assistant and model family made by xAI. Groq is a separate company that builds LPU inference chips and runs GroqCloud, an API for other labs' open models.

### Who owns Groq?

Groq LLC is a private company. Its recent rounds were led by Disruptive and Infinitum, with Alex Davis of Disruptive as chairman and Adam Winter as CEO, and it has not published ownership shares. Its December 2025 deal with Nvidia was a non-exclusive licensing agreement for its technology.

### Is Grok free?

Yes, with limits. As of November 16, 2026, The X Subscription has a free plan with limited Grok usage, and paid plans cost $8, $30, $100 or $200 a month.

### Can I use Grok models on GroqCloud?

No. As of November 16, 2026, GroqCloud's model list has no Grok model. Grok models are available through the xAI API at api.x.ai.

## Sources

- [Pricing (The X Subscription)](https://docs.x.ai/x-subscription/pricing), xAI
- [API pricing](https://docs.x.ai/developers/pricing), xAI
- [What is The X Subscription?](https://docs.x.ai/x-subscription/overview), xAI
- [Grok](https://grok.com/), xAI
- [GroqCloud models](https://console.groq.com/docs/models), Groq
- [GroqCloud rate limits](https://console.groq.com/docs/rate-limits), Groq
- [GroqCloud overview](https://console.groq.com/docs/overview), Groq
- [Groq platform](https://groq.com/platform), Groq
- [Groq and Nvidia enter non-exclusive inference technology licensing agreement](https://groq.com/newsroom/groq-and-nvidia-enter-non-exclusive-inference-technology-licensing-agreement-to-accelerate-ai-inference-at-global-scale), Groq
- [Groq raises $650M to scale its AI inference cloud business](https://groq.com/newsroom/groq-raises-usd650m-to-scale-its-ai-inference-cloud-business), Groq
- [Groq becomes an NVIDIA Cloud Partner](https://groq.com/newsroom/groq-becomes-an-nvidia-cloud-partner), Groq
- [Groq closes $350 million Series A](https://groq.com/newsroom/groq-closes-usd350-million-series-a-building-the-world-s-leading-ai-inference-cloud), Groq
