Grok AI Pricing (July 2026): What You Pay, What You Get & Whether It’s Worth It

TL;DR: Grok AI has eight pricing tiers ranging from $0 to $300/month. The free plan gives you roughly 10 prompts every two hours. SuperGrok at $30/month is the best value for most users, offering full model access, image generation, and 128K token context. SuperGrok Heavy at $300/month unlocks multi-agent mode and maximum compute priority. The API starts at $1.25 per million input tokens for the grok-4.3 model.

If you’ve landed here, you’re probably staring at Grok’s pricing page feeling mildly confused. Six consumer tiers. Two business plans. An API with its own pricing table. And a gap between the cheapest and most expensive plan that spans $292 per month.

That’s a lot to unpack.

This guide breaks down every single Grok AI pricing tier clearly, what you actually get, what you don’t, who each plan is right for, and where the value genuinely exists. No fluff. No guesswork. Just the real picture so you can make a confident decision.

Prices are verified as of July 2026, sourced from x.ai’s official pricing and API pages.

Is Grok Really Free?

At one point, Grok was completely free to use. But now, the free plan only gives you access to Fast Mode. You can still chat with Grok, but there are usage limits. You can also generate a few images, usually around two or three, before hitting the limit. Apart from that, most of Grok’s features are now locked behind paid plans.

The Full Grok Pricing Breakdown (2026)

Here’s every active plan, side by side:

SuperGrok Lite launched on March 25, 2026. SuperGrok Heavy has been available since July 9, 2025. Neither is widely covered by most pricing guides. So let’s go deeper on all of them.

What Do You Get with Each Plan?

Free ($0/month)

Already covered above. Use it to try Grok. Don’t rely on it for real work.

X Premium ($8/month)

This is Grok bundled with an X subscription. You get moderately higher query limits than the free tier, a blue checkmark on X, and access to Grok inside the X platform. What you don’t get: Grok 4 Heavy, DeepSearch, full image generation, or the Companions feature.

The honest framing: if you already pay for X Premium and want a bit more AI access, this is fine. If your goal is specifically to get more out of Grok as an AI tool, the math works out better with SuperGrok Lite at $10.

SuperGrok Lite ($10/month)

The cheapest standalone Grok subscription. You get access to the Grok 3.5 model, basic Imagine image generation, and approximately 15 videos per day at 480p resolution with a 6-second maximum duration. Chat sessions are about 2x longer than the free tier.

SuperGrok Lite exists for people who want to try a paid plan without the $30/month commitment. It’s a stepping stone, not a destination.

SuperGrok ($30/month)

This is the plan most people should be on.

You get:

• Grok 4.5 model (xAI’s current flagship as of July 2026, launched July 8, 2026)
• DeepSearch – Grok’s deep research mode
• Big Brain Mode – extended reasoning for complex tasks
• Full Imagine – image and video generation
• 128K token context window – meaning you can send and receive long, detailed conversations without losing context
• Higher rate limits across all features
• Expert feature
• SOC 2 Type I & II compliance
• Connectors

One important note: SuperGrok users receive Grok 4.3 and 4.5 access via staged rollout. More on what that actually means in a moment.

Annual billing is available at $300/year, which saves you $60 compared to paying month-to-month.

X Premium+ ($40/month)

X Premium+ bundles full Grok access with the premium X social experience: ad-free browsing, higher post visibility, and priority support from X. On the AI side, you get Grok 4 access and higher message throughput compared to X Premium.

The comparison to SuperGrok is straightforward. SuperGrok gives you more AI features (DeepSearch, Big Brain Mode, Imagine, 128K context) for $10 less per month. X Premium+ wins only if you genuinely value both the AI and the X platform experience. If you’re paying $40 just for Grok, you’re overpaying.

SuperGrok Heavy ($300/month)

The top tier. And the one that raises the most eyebrows.

SuperGrok Heavy unlocks three things that no other consumer tier offers:

1. Multi-agent mode – Grok 4 Heavy can run multiple AI agents in parallel, tackling complex tasks simultaneously rather than sequentially. Independent reporting describes up to 16 parallel agents, though xAI doesn’t confirm the exact number.
2. 256K- 428K token context window – roughly double or more what standard SuperGrok offers. Useful for feeding in large documents, codebases, or research datasets in a single session.
3. Confirmed Grok 4.3 access from day one – not a staged rollout. You’re always on the current flagship.

You also get maximum compute priority, meaning your queries jump the queue.

Heavy mode has posted strong benchmark scores: HLE 50.7% (text-only subset), AIME 2025 100%, and GPQA Diamond 88.9%, per xAI’s published results.

Is it worth $300/month? For most people, no. For researchers, quantitative analysts, and teams doing compute-heavy work like large-scale document analysis or complex technical reasoning – it’s a different calculation. More on that below.

Business ($30/seat/month)

Everything in SuperGrok, plus:

• Team seat management and consolidated billing
• Google Drive integration
• Role-based access control
• Your data is NOT used to train Grok models – this is a firm guarantee that free and consumer plans don’t provide
• SOC 2 Type I & II compliance

At $30/seat, Grok Business is notably affordable compared to competing team AI products. It’s a strong option for small teams that want full Grok features with admin controls and data privacy guarantees.

Enterprise (Custom Pricing)

Unlimited users, custom SSO and SCIM (directory sync), advanced audit controls, customer-managed encryption keys (CMEK) via Grok’s Enterprise Vault, dedicated data plane, custom data retention, and dedicated onboarding support. Contact xAI’s sales team at [email protected] or via x.ai/contact-sales for pricing.

The Model Transparency Gap Nobody Talks About

Here’s the uncomfortable truth: on most consumer tiers, you can’t always tell which model version is actually processing your queries.

xAI uses something called Auto Mode, which dynamically routes your query to a model variant based on perceived complexity. Two SuperGrok subscribers can send the same prompt at the same time and hit different model versions during a staged rollout. No UI indicator tells you which model responded.

This doesn’t mean you’re being misled. It’s how staged rollouts work across the industry. But it does mean that if model version precision matters to your workflow, the consumer apps aren’t the right tool. The API – where you can specify a dated model ID like grok-4.3 – is the only way to know exactly what processed your query.

SuperGrok Heavy is the one exception: it guarantees confirmed access to the current flagship model from day one, no staging uncertainty.

SuperGrok vs. SuperGrok Heavy: Is the $270/Month Gap Worth It?

The short answer: for most people, no. For the right use case, yes.

The difference between these two plans isn’t about features like image generation or DeepSearch; both plans have those. The gap is pure compute. SuperGrok Heavy gives you parallel multi-agent execution, a much larger context window and priority queuing. Standard SuperGrok runs on single-agent mode with a 128K context limit.

Choose SuperGrok ($30/month) if you use Grok daily for research, writing, analysis or creative work. Full model access, DeepSearch and Imagine will cover 95% of what everyday power users need.

Choose SuperGrok Heavy ($300/month) if you’re doing compute-intensive tasks like large-scale research synthesis, feeding thousands of pages of documents into a single session, or multi-step technical reasoning where parallel agents meaningfully speed up your output. The math only works if the additional throughput translates to real productivity gains.

If you’re unsure which one you need, start with SuperGrok. You’ll know quickly if you’re running into its limits.

Grok API Pricing for Developers

If you’re building on top of Grok rather than using it directly, the API is your path. It’s OpenAI-compatible, so migrating from OpenAI is as simple as changing your base URL and swapping in an xAI API key.

Current API Models and Rates

New API accounts receive $25 in free trial credits to get started.

Imagine API (Image and Video)

• Image (quality): $0.05–$0.07 per image at 1K–2K resolution
• Image (standard): $0.02 per image at 1K–2K resolution
• Video: $0.05–$0.07 per second at 480p–720p

Voice API

• Speech-to-Speech: $0.05/min ($3.00/hr)
• Text-to-Speech: $15.00 per 1M characters
• Speech-to-Text: $0.10/hr (REST), $0.20/hr (Streaming)

Tool Calls (Billed Separately)

A detail that catches a lot of developers off guard: built-in tools carry separate fees on top of token costs.

• Web Search / X Search / Code Execution: $5.00 per 1,000 calls
• File Attachments: $10.00 per 1,000 calls
• Collections Search (RAG): $2.50 per 1,000 calls

A single research query can trigger 3 to 5 tool calls. That adds roughly $0.015 to $0.025 in tool fees per query, according to xAI documentation. For high-volume applications, design your prompts to minimize unnecessary tool invocations.

Storage pricing (file storage: $0.025/GiB/day; collection storage: $0.10/GiB/day) has been in effect since April 2026.

One important heads-up: Several older model IDs (grok-4, grok-4-fast, grok-4.1, grok-4.1-fast, grok-3, and others) were retired on May 15, 2026. Requests to those slugs don’t fail – they silently redirect to grok-4.3 and bill at grok-4.3 rates. If you built cost estimates on the old model pricing, recheck your projections. The batch API offers a 20–50% discount for async workloads processed within a 24-hour window.

How Grok Pricing Compares to ChatGPT, Claude and Gemini

Here’s how the main AI platforms stack up in 2026:

A few things stand out here.

Gemini is the cheapest paid entry at $4.99/month. ChatGPT Plus at $20/month remains a popular middle-ground. Claude’s entry point at $20 puts it at the same level as ChatGPT Plus.

Grok’s $30/month SuperGrok plan is slightly pricier than ChatGPT Plus, but it comes with something neither ChatGPT nor Claude offer at that price point: real-time access to live X (Twitter) data. That’s a meaningful differentiator for anyone tracking news, financial signals or social sentiment.

Grok’s Business plan at $30/seat is aggressively priced compared to similar team plans from competitors. And SuperGrok Heavy at $300/month has no direct equivalent in ChatGPT or Claude’s consumer lineup – neither platform currently offers a consumer-facing multi-agent compute tier at that level.

Claude’s API pricing is significantly higher than Grok’s – Claude Sonnet 4.6 runs $3/$15 per 1M tokens input/output versus Grok’s grok-4.3 at $1.25/$2.50. For developers building at scale, that difference compounds fast.

One note on data privacy: if you’re on the free tier or any consumer plan, xAI may use your conversations to improve its models. Only Business and Enterprise plans come with a hard guarantee that your data won’t be used for training. If you’re working with sensitive information, that matters.

Grok Pricing History: What’s Changed in the Last 12 Months

Grok’s pricing has moved a lot in a short time. Here’s the key timeline:

• July 9, 2025 – SuperGrok Heavy launched at $300/month. The first premium compute tier in Grok’s history.
• September 19, 2025 – Grok 4 Fast introduced at $0.20/$0.50 per 1M tokens (API).
• November 17, 2025 – Grok 4.1 and Grok 4.1 Fast added to the API.
• March 10, 2026 – Grok 4.20 multi-agent model reached API general availability.
• March 25, 2026 – SuperGrok Lite launched at $10/month, adding an affordable standalone entry point.
• May 1, 2026 – Grok 4.3 reached API GA at $1.25/$2.50 per 1M tokens.
• May 15, 2026 – Eight older API models retired and redirected to grok-4.3 billing rates.
• May 29, 2026 – Grok Build 0.1 coding model entered public beta at $1.00/$2.00 per 1M tokens.
• Q2 2026 – Text-to-Speech API pricing jumped from $4.20 to $15.00 per 1M characters, a steep increase for developers relying on voice features.
• July 8, 2026 – Grok 4.5 launched as the new flagship model, now included in  SuperGrok.

The overall pattern: xAI is consolidating its API into fewer, more capable models and simplifying the consumer tier structure. Prices on the API have generally trended downward with each model generation for equivalent capability, except for the TTS increase, which went the other direction.

If you built API integrations using older model slugs, check your billing. Those retired model IDs now route to grok-4.3 rates whether you asked for that or not.

Frequently Asked Questions

The Final Verdict: Which Grok Plan Is Actually Worth It?

The free plan is good for testing Grok, but it isn’t enough for regular use.

SuperGrok at $30/month is the best choice for most users. It includes the latest model, image and video generation, DeepSearch, and enough context for writing, research, coding and creative work.

X Premium and Premium+ only make sense if you already use X regularly.

SuperGrok Lite is a cheaper option for users who need longer chats and basic image generation.

SuperGrok Heavy is built for heavy workloads and advanced users. Most people won’t need it.

For developers, the API pricing is flexible, and the $25 free credits make it easy to test before scaling.

The main rule is simple: pick the plan that matches your usage and avoid paying for features you won’t use.