Grok (xAI)
Grok is xAI's assistant, distinguished by real-time grounding in X and the live web: it answers about what is happening now, not what the training data remembers. The model line has pushed to the benchmark front on reasoning, with long context, deep research modes and document generation arriving through successive releases.
The tier range is unusually wide, from a rate-limited free tier through mid-price plans to heavy top tiers, with voice features and agent tooling growing quickly around the core.
The brand-safety question is the standing caveat: its permissiveness record makes it a personal-productivity choice more comfortably than a corporate-default one.
- Cost
- Freemium (Free tier + paid plans)
- Ease
- Model
- Hosted service
- Checked
- August 2026
Prices, plans and model versions change fast: this is a mid-2026 snapshot; check the tool's official site for the latest.
Best for
- Real-time answers grounded in X and the live web
- Benchmark-front reasoning on the top models
- Tracking breaking topics as they develop
- A wide tier range from free to heavy use
- Users already living inside the X ecosystem
Less suited to
Grok's permissiveness record, including a documented image-generation safety scandal, makes it hard to defend as a corporate default where brand safety and regulated data are in play.
Its ecosystem is also younger: workplace integrations and governance tooling trail the enterprise incumbents.
Costs & data, in short
Freemium, with a free tier that is genuinely usable, a personal subscription above it, a per-seat business tier at the same headline rate, and a quote-only enterprise tier. The bundling is the thing to check rather than the number: access has moved between the assistant's own subscription and the wider X subscription more than once.
Its distinguishing feature is live grounding in X and the open web, so prompts travel to xAI and answers are shaped by public posts; treat anything confidential as unsuitable, and expect the provenance of a real-time answer to be a moving target rather than a citable source.
Plans
| Free | Free |
|---|---|
| SuperGrok | $30per month |
| Business | $30per user, per month |
| Enterprise | Price on applicationno list price published |
Prices as of August 2026. Prices and plans change regularly. Check with the provider before you buy.
In practice
How Grok (xAI) is used, area by area.
Jobs
Getting started
Grok is free for anyone on X and in its own apps
Grok is free for anyone on X and in its own apps, within limits, which makes it an easy first assistant to simply try. Its clearest newcomer strength is currency: it reads live public posts and cites web sources, so it is more up to date than assistants bound to a fixed training cutoff, and following the citations teaches the checking habit early, which is the habit worth forming first. Ask about something that happened this week and see the difference. The tier range runs unusually wide, from free to heavy use, so growing into it costs nothing up front. The standing caveat travels with it: the permissiveness record makes it easier to defend as a personal-productivity choice than as a company default, so for workplace or sensitive use the more governed assistants are the safer start, and its looser, opinionated tone is a deliberate choice some organisations treat as risk.
Example tasks
- Ask about things happening right now with live grounding
- Follow a breaking story with sourced context from X
- Draft and rewrite everyday text
- Try a deep research mode on a topic you know
- Compare its answers with your usual assistant before switching
Limits
You want the most neutral, predictable tone or the strictest workplace content controls. Grok's looser, more opinionated personality is a deliberate choice that some organisations treat as a governance or brand-safety risk; for buttoned-up business use, ChatGPT, Claude or a governed Copilot are safer defaults.
Compares
| vs | Pick Grok (xAI) when | Pick the other when |
|---|---|---|
| ChatGPTFull comparison → | Grok's newcomer edge is currency and a free tier for anyone already on X, reading live public posts and citing web sources so it is more up to date than an assistant bound to a fixed training cutoff | the broader ecosystem and the shared know-how around it matter more to a beginner than knowing what happened this morning |
| ClaudeFull comparison → | Grok trades predictability for personality and live data, which suits someone following a fast-moving story | careful reasoning and a steady tone are what a beginner needs while they learn to judge an answer |
| Google GeminiFull comparison → | Grok needs no ecosystem and asks only that you already use X | your work runs on Workspace and integration into Gmail and Docs is the friction worth removing first |
| PerplexityFull comparison → | Grok reads the live X feed alongside web sources, which is a narrower and more current corpus | the habit worth forming is checking every answer against numbered sources across the wider web |
| KimiFull comparison → | Grok has a beginner follow its citations, which teaches the checking habit early and is the habit worth forming first | the free tier should let a newcomer learn every surface before anything is spent |
| Mistral VibeFull comparison → | Grok invites a newcomer to compare its answers with their usual assistant before switching, which is a fair test to offer | responses coming back quickly matter more than they sound while you experiment your way into new habits |
Tasks
Search & knowledge retrieval
Grok's structural advantage in retrieval is native real-time access to X
Grok's structural advantage in retrieval is native real-time access to X. Competitors cannot replicate that access, and a DeepSearch mode pulls and cites sources rather than answering from memory alone. For fast-moving topics, breaking news, developing stories, what people are saying right now, a fixed-cutoff model guesses or hedges where Grok reads the live feed, and the citations keep the answers checkable. Anyone whose questions concern the current moment gets something genuinely different here. The sourcing is also the caveat: X-sourced answers inherit X's biases and noise, broader and more neutral web sourcing is the citation-first engines' ground, and retrieval over your own documents belongs to the grounded corpus tools rather than a live-feed assistant.
Example tasks
- Answer time-sensitive questions from the live web and X
- Track a developing story as sources appear
- Run deep research on current topics
- Gauge sentiment around a topic on X
- Verify freshness-critical claims before publishing
Limits
You need retrieval grounded in your own documents (NotebookLM or a RAG setup fits better) or broader, more neutral sourcing (Perplexity cites the wider web). X-sourced answers also inherit X's biases and noise.
Compares
| vs | Pick Grok (xAI) when | Pick the other when |
|---|---|---|
| PerplexityFull comparison → | Grok is stronger on live X specifically, reading the feed directly where a fixed-cutoff model would guess or hedge | the research wants broader, more neutral web sourcing with numbered citations |
| ChatGPTFull comparison → | Grok's differentiator is native real-time access to X, with a DeepSearch mode that pulls and cites sources rather than answering from memory | the question is not about the current moment and synthesis across your own material matters more |
| Google GeminiFull comparison → | Grok reads the live X feed directly rather than grounding in a search index, which is what fast-moving questions turn on | reach across the public web and your own Workspace matters more than one live source |
| ClaudeFull comparison → | Grok answers about the present, where a model working from a fixed corpus would hedge | the work is careful synthesis across documents you supply rather than anything happening right now |
| Microsoft Copilot | Grok's corpus is the public live feed | the question is about your own organisation's email, chats, meetings and files |
| Notion AI | Grok covers what is being said publicly right now | the answer is institutional memory sitting in the team's own pages and connected tools |
| KimiFull comparison → | Grok tracks a developing story as the sources appear and will gauge the sentiment around a topic on X, which is a reading of the present rather than a report on it | the question is open enough to need a deep research run behind the answer |
| Mistral VibeFull comparison → | Grok verifies freshness-critical claims before they are published, which is a question about the last few hours rather than the last few years | the hosting jurisdiction has to be defensible from the start |
Writing & research
Currency is Grok's case in writing and research
Currency is Grok's case in writing and research. Everyday drafting, rewriting and research jobs run competently, and its reasoning modes plus live search help precisely where general writers stumble: tasks that need current facts or step-by-step working, with sources cited rather than remembered. Writers whose material moves fast, and who accept a lighter, more opinionated default tone, get an assistant that knows what happened this morning. Polish and predictability are the trade: publication-grade prose and careful long-form reasoning are more commonly the measured assistant's ground, and regulated writing that needs predictable tone and strict content controls sits outside its comfort zone, since the loose register is a feature, not an accident.
Example tasks
- Research fast-moving topics with current grounding
- Draft commentary that references what happened today
- Summarise live discussion around an event
- Generate documents from research on recent developments
- Cross-check dated claims in older drafts
Limits
Publication-grade prose or careful long-form reasoning (Claude is the more common pick), or regulated writing needing predictable tone and strict controls.
Compares
| vs | Pick Grok (xAI) when | Pick the other when |
|---|---|---|
| ClaudeFull comparison → | Grok trades prose polish and careful long-form reasoning for live data and a lighter, more opinionated register | the piece is publication-grade or the reasoning has to hold over length |
| ChatGPTFull comparison → | Grok is close on general writing and separated by what it can reach, with live search and reasoning modes for tasks needing current facts | the ecosystem and the range across research, drafting and revision matter more |
| Google GeminiFull comparison → | Grok cites what it found this morning rather than holding a large fixed corpus | the writing rests on book-length sources and the drafting belongs in Google's tools |
| PerplexityFull comparison → | Grok writes as well as sources, with reasoning modes attached to the same session | the deliverable is the evidence itself, sourced and bibliographied, and the prose happens elsewhere |
| KimiFull comparison → | Grok cross-checks the dated claims in an older draft and summarises the live discussion around an event, which is the work that goes stale fastest | the deep research surface should do the digging before there is a draft at all |
| Mistral VibeFull comparison → | Grok generates documents from research on recent developments, where the value of the material decays by the week | the work is everyday drafting and research and a European vendor is a procurement advantage |
Where to start
Not sure what to adopt first?
Five quick questions about your job, task and constraints. We'll suggest your top three tools, plus the one to try first.
Alternatives
Same category, different strengths.
Common questions
Is Grok (xAI) free?
There's a free tier to start; paid plans add capacity and features.
Where does Grok (xAI) fit best?
Grok (xAI) fits best in Getting started and Search & knowledge retrieval; see its practice notes for how.
Before sharing confidential or personal data, check this tool's data-governance and training policies. They differ between providers and can change.
Last checked: August 2026