Grok
xAI's chatbot on Grok 4.5, with live X data, DeepSearch, Imagine Image 2.0 and Video 1.5 generation, plus the Grok Build coding agent.
Updated 2026-08-10
Yes, within limits. Grok runs a free tier with paid plans above it.
Listed pricing: Free tier + SuperGrok $30/mo + Heavy $300/mo (or X Premium+ $40/mo).
Overview
Grok is xAI's flagship assistant, and the 2026 version of it is less a chatbot than a bundle. The flagship model has been Grok 4.5 since 8 July 2026 — the same Opus-class model xAI shipped into Cursor and Perplexity, which is the most useful third-party signal of where its reasoning actually lands, since xAI publishes little that can be independently checked. The structural advantage hasn't changed: native access to real-time X posts, trends and public discourse, which no rival chatbot can reach. DeepSearch builds directly on that, decomposing a question into sub-queries across the live web and X and returning sourced synthesis.
What changed since the spring is almost everything around the model. Aurora is no longer the image engine — Quality Mode in Grok Imagine now runs Imagine Image 2.0, announced 7 August 2026, which adds object-level editing, multi-image compositing, background removal and reframing inside the same model that made the image, and a free Grok account reaches it under rate limits xAI does not publish. Grok Imagine Video 1.5 handles image-to-video with native audio. Grok Build, xAI's terminal coding agent, left beta at v1.0 on 7 August 2026 with MCP tool support, sandboxed execution, codebase restore and plan-review-before-edit — and it is included in the same SuperGrok subscription rather than sold as a separate product. Build Mode inside the consumer app generates shareable websites, apps and games.
The honest take: the pricing has not moved — free tier, SuperGrok at $30/mo, SuperGrok Heavy at $300/mo, with X Premium+ at $40/mo as a side door — so the same money buys substantially more than it did in March. What Grok still does not have is computer use, which now ships inside the ChatGPT, Claude and Gemini subscriptions most buyers already pay for. And the release cadence is its own tax: Musk put Grok 4.6 roughly two weeks out on 28 July with 4.7 behind it, so the flagship version named here may already be one release stale by the time you read it.
Is Grok free?
Yes, within limits. Grok runs a free tier with paid plans above it.
What the free tier covers: Genuinely usable: free accounts get Grok chat, live X data and Imagine Image 2.0 including Quality Mode. The catch is that xAI does not publish the daily limits, and they tighten under load.
Listed pricing: Free tier + SuperGrok $30/mo + Heavy $300/mo (or X Premium+ $40/mo).
Pricing verified . That is the date the plans were last re-checked against the vendor's own pages, not today's date. Confirm on the official site before you pay.
What the free tier leaves out
Read straight off the plan list below. Vendors move features between tiers, so check the current split before you pay.
- SuperGrok $30/mo Higher limits, Grok 4.5 with extended thinking, DeepSearch, image and video generation, and full Grok Build access. Specific rate limits are not publicly documented.
- X Premium+ $40/mo Grok access plus Grok Build, bundled with X platform features. A separate door into the same tooling for people already paying for X.
- SuperGrok Heavy $300/mo Highest usage ceilings across all Grok products, multi-agent collaboration and earliest access to new model releases.
Grok pricing
| Plan | Price | What's included |
|---|---|---|
| Free | $0 | Grok chat, real-time X data and Imagine image generation including Quality Mode. Rate limits are not published as a fixed number and tighten under load. |
| SuperGrok | $30/mo | Higher limits, Grok 4.5 with extended thinking, DeepSearch, image and video generation, and full Grok Build access. Specific rate limits are not publicly documented. |
| X Premium+ | $40/mo | Grok access plus Grok Build, bundled with X platform features. A separate door into the same tooling for people already paying for X. |
| SuperGrok Heavy | $300/mo | Highest usage ceilings across all Grok products, multi-agent collaboration and earliest access to new model releases. |
Grok chat, real-time X data and Imagine image generation including Quality Mode. Rate limits are not published as a fixed number and tighten under load.
Higher limits, Grok 4.5 with extended thinking, DeepSearch, image and video generation, and full Grok Build access. Specific rate limits are not publicly documented.
Grok access plus Grok Build, bundled with X platform features. A separate door into the same tooling for people already paying for X.
Highest usage ceilings across all Grok products, multi-agent collaboration and earliest access to new model releases.
Pricing verified . That is the date the plans were last re-checked against the vendor's own pages, not today's date. Confirm on the official site before you pay.
Is Grok worth it?
Worth it for Journalists and analysts tracking breaking news and sentiment on X who also want a frontier chat model, a terminal coding agent, and image plus video generation inside a single $30/mo subscription.
You can test that on the free tier before paying anything. The recorded trade-offs are listed below, and any one of them can settle the question on its own.
The 8.9/10 AI Score is an editorial read of published capability, price and shipping pace. Nobody here has hands-on hours with Grok. How we verify.
Worth it if
The strengths recorded against this entry.
- Real-time X/Twitter data integration remains unmatched by any competing chatbot, at any tier
- One $30/mo subscription bundles a frontier chat model, a v1.0 terminal coding agent, image generation with in-model editing, and image-to-video with audio
- Free tier reaches Imagine Image 2.0 Quality Mode — the same engine paid users get, which most rivals gate behind a subscription
- Grok 4.5's adoption by Cursor and Perplexity is real external evidence of frontier-level coding ability, not just a vendor claim
- More permissive content policies than competitors, with fewer refusals on political or edgy topics
Not worth it if
Any one of these blocks your use case.
- No computer use, which now ships inside the ChatGPT, Claude and Gemini subscriptions most buyers already hold
- xAI publishes very little that can be verified: no free-tier rate limits, no stated context-window size, few independent benchmarks
- Release churn is disorienting — 4.5 landed in July with 4.6 and 4.7 signaled within weeks, so documentation and workflows go stale fast
- Heavy reliance on X data can skew answers toward Twitter-centric framing, and the $300/mo Heavy tier is hard to justify without extreme rate-limit needs
What sets Grok apart
- Native real-time X/Twitter data access — a structural advantage no other chatbot can replicate
- One consumer subscription includes a v1.0 terminal coding agent (Grok Build), image generation with in-model editing, and image-to-video with native audio
- Free accounts reach the same Imagine Image 2.0 Quality Mode engine as paying users, where Midjourney has no free tier at all
- Two separate paid doors to the same tooling — SuperGrok at $30/mo or X Premium+ at $40/mo bundled with the X platform
Key features
Grok 4.5 Reasoning
The flagship model since 8 July 2026, described by xAI as Opus-class and adopted by Cursor and Perplexity for their own coding surfaces — the closest thing to external validation available, since xAI publishes few independent benchmark comparisons.
Real-time X Data + DeepSearch
Native access to live X posts, trends and public conversation, with DeepSearch decomposing a question into sub-queries across the web and X and returning sourced synthesis. Still the one capability no competing chatbot can match at any price.
Imagine Image 2.0 and Video 1.5
Image generation now runs Imagine Image 2.0 (7 August 2026), replacing Aurora and adding object-level edits, compositing, background removal and reframing inside the model. Grok Imagine Video 1.5 covers image-to-video with native audio. Both reachable on a free account under unpublished limits.
Grok Build and Build Mode
Grok Build, the terminal coding agent, reached v1.0 on 7 August 2026 with MCP tools, sandboxed runs, session forking and full-script permission prompts — included in SuperGrok, not billed separately. Build Mode in the consumer app generates shareable websites, apps and games.
How it compares
| Tool | Best for | Pricing | Score |
|---|---|---|---|
| Grok | Journalists and analysts tracking breaking news and sentiment on X who also want a frontier chat model, a terminal coding agent, and image plus video generation inside a single $30/mo subscription. | Free tier + SuperGrok $30/mo + Heavy $300/mo (or X Premium+ $40/mo) | 8.9/10 |
| Claude vs Claude → | Developers and technical teams who want a coding agent running in their own repo, terminal and IDE rather than a hosted editor, plus a chat assistant strong enough for long-form writing and 1M-token codebase analysis. | Free + Pro $17-20/mo + Max from $100/mo + Team $20-25/seat | 9.4/10 |
| ChatGPT vs ChatGPT → | Generalists and small teams who want one $20 subscription covering reasoning, full-duplex voice, image and video generation, Deep Research and a phone-controllable coding agent instead of stitching four separate tools together. | Free + Go (low-cost, regional) + Plus $20/mo + Pro $200/mo; Business and Enterprise above | 9.3/10 |
| GPT-5.6 Sol vs GPT-5.6 Sol → | Engineering teams building long-horizon coding or tool-use agents who need a frontier reasoning tier for the hard steps and can route easier calls to Terra or Luna on the same API key. | API usage-based: Sol $5/$30 per 1M tokens; cheaper Terra and Luna tiers cut July 30 | 9.3/10 |
Learn to use Grok
Step-by-step, each one stamped with the date it was last checked against the docs.
Compare head-to-head
Related reading
Grok 4.6 Release Date: Two Weeks Out, Per Musk
Elon Musk put Grok 4.6 about two weeks out on July 28, with 4.7 close behind. What that timeline changes for anyone already building on Grok 4.5.
Grok 4.5 vs GPT-5.6 Sol: What the Benchmarks Say
A sourced head-to-head on Grok 4.5 and GPT-5.6 Sol: coding scores, intelligence index, pricing, and what's still unverified.
Grok 4.5 Drops: xAI's Opus-Class Coding Model
xAI launched Grok 4.5 on July 8 as an "Opus-class" model for coding and agents—now default in Grok Build and live in Cursor and Perplexity.
Ready to try Grok?
Head to the official site to start with Grok — pricing and plans are listed above.
Visit Grok