Grok
Identity card
The company
- Founded
- 2023
- Head office
- Palo Alto, California, United States
- Chief executive
- Elon Musk
The product
- Flagship model
- Grok 4.7
- Subscription
- $30 per month · Premium, price in the United States
- API
- 2 $ / 6 $ input / output, per million tokens
The data
- Law
- United States
Usage
- Web traffic, Sept. 2026
- 2.3 %
Key points
Grok, the assistant from SpaceX Super Intelligence (formerly xAI) built into the X network, relies on real-time news and on the freedom of tone claimed by Elon Musk. That freedom produced millions of non-consensual sexualised images in early 2026, and the company publishes almost nothing about its models and their footprint.
- Following the news on X
- Video
- Teenagers
- Professional use
- Sobriety
Source: Similarweb, share of visits to generative AI sites, worldwide; records published by Similarweb and repeated by the press.
Yes for following the news as it circulates on X. Look elsewhere for family, professional or environmentally conscious use (Claude, Gemini).
Strengths
- Built into the social network XGrok reads X posts in real time; its AI features had nearly 117 million monthly users in March 2026, according to SpaceX.Source: Pulse 2.0, citing SpaceX, 27 May 2026 ↗
- A widely used video generatorGrok Imagine produced 1.245 billion videos in 30 days, according to xAI in February 2026.Source: Gigazine, 3 February 2026 ↗
- Backed by SpaceXAbsorbed by SpaceX in February 2026, xAI is part of a group valued at more than 1,000 billion dollars.Source: Built In, 3 February 2026 ↗
Weak points
- Non-consensual sexualised imagesThe CCDH counted nearly 3 million sexualised images produced with Grok in 11 days, in January 2026; the Irish data protection authority has been investigating since February.Source: The Star, citing the CCDH, 23 January 2026 ↗ ; RTÉ, 17 February 2026 ↗
- Antisemitic remarks in 2025In July 2025, after an update, Grok made antisemitic remarks, going so far as to call itself "MechaHitler"; the European Commission said it was taking the matter very seriously.Source: The Times of Israel, July 2025 ↗
- Last in the Stanford transparency indexThe Stanford transparency index gives xAI 14 out of 100, joint last; the company publishes no energy figure, no environmental report and no target.Source: Stanford, Foundation Model Transparency Index, December 2025 ↗
The company SpaceX Super Intelligence
SpaceX Super Intelligence entry →Company history
xAI was founded in March 2023 by Elon Musk, a co-founder of OpenAI which he left, with 11 researchers. It unveiled Grok in November 2023, at first reserved for X subscribers.
The story continues
In March 2025, xAI absorbs X, the former Twitter; in February 2026, SpaceX absorbs xAI. Grok aims to be less restricted than its competitors, and its missteps have drawn several investigations in Europe.
xAI took the name SpaceXAI in summer 2026; in October 2026, Elon Musk announced the name SpaceXSI, and the Grok privacy policy of 27/10/2026 names the operator as SpaceX Super Intelligence LLC.
Sources: Wikipedia, 7 November 2026 ↗ ; Built In, 3 February 2026 ↗ ; Siècle Digital, 6 October 2026 ↗ ; SpaceX Super Intelligence (formerly xAI), privacy policy (27/10/2026), 9 November 2026 ↗
Key dates
- 9 March 2023Founded by Elon Musk and 11 researchers↗
- 4 November 2023Unveiling of Grok↗
- 28 March 2025xAI absorbs X↗
- 2 February 2026SpaceX absorbs xAI↗
- 17 February 2026Irish authority investigation into the sexualised images↗
- 27 October 2026xAI, which became SpaceXAI in the summer, is now called SpaceX Super Intelligence (SpaceXSI)↗
Leadership
- Founder
- Elon Musk ↗
Key figures
- Revenue of SpaceX's AI division, 1st quarter 2026
- 818 million dollarsReuters (The Star), citing the SpaceX prospectus, 20 May 2026 ↗
- Losses of SpaceX's AI division, 1st quarter 2026
- 2.47 billion dollarsReuters (The Star), citing the SpaceX prospectus, 20 May 2026 ↗
Grok plans
- FreeFree0 $
- PlusStandard8 $
- PremiumStandard30 $
- SuperPremium100 $
- UltraPremium200 $
SpaceX Super Intelligence does not publish a price in euros: the amount paid in France depends on the exchange rate and VAT. Pricing pages, up to date on 9 November 2026.
Content of the 5 plans, prices in France and the United States
Price in dollars only. When only the price in dollars is shown, the company does not publish a price in euros: the amount paid in France depends on the exchange rate and VAT.
VibeMistral Large1.50 $0.50 $
PerplexitySonar2.50 $0.25 $
DeepSeekV4-Pro3.96 $1.32 $
Meta AIMuse Spark 1.14.25 $1.25 $
QwenQwen3.8-Max6 $2 $- GrokGrok 4.76 $2 $4 times the output price of Vibe
GeminiGemini 4 Argon10 $2 $
ClaudeOpus 5.520 $4 $
ChatGPTGPT-6 Astra50 $10 $
Copilot: no API of its own. Flagship model of each company, standard rate, up to date as of 9 November 2026.
Prices up to date as of 9 November 2026, read on the companies' pages, or from 2 concordant sources; the date of each record is in the page's price file Methodology.
Grok models
- Company
- SpaceX Super Intelligence
Links read on 10/11/2026 on the company's pages, press as a fallback. Rank: Epoch AI ECI index (CC BY 4.0). Performance of AI models
Data management
How SpaceX Super Intelligence handles conversations
- Training models on conversationsYes, unless you opt out: questions, uploaded files and Grok's replies may be used to train SpaceX Super Intelligence models. "Private chat" is never used, nor are business accounts.Source: SpaceX Super Intelligence (formerly xAI), Grok consumer FAQ, 9 November 2026 ↗
- Opting out of trainingIn the Grok app: Settings, Data Controls, untick "Improve the model"; on grok.com: Settings, Data. Opting out applies to later conversations, and feedback given on a reply may still be used. In X, the setting is described by the X help centre.Source: SpaceX Super Intelligence (formerly xAI), Grok consumer FAQ, 9 November 2026 ↗
- Retention periodDeleted conversations, or those held in private chat, are erased within 30 days, unless required by law or for security.Source: SpaceX Super Intelligence (formerly xAI), privacy policy (27/10/2026), 9 November 2026 ↗
- Applicable lawUS law: SpaceX Super Intelligence is a United States company, subject in particular to the Cloud Act, even when the servers are in Europe.Source: Guide d'usage de l'IA pour les agents publics de l'État (DINUM), June 2026 ↗
Score on this criterion: 17 out of 100, 9th out of 10. Training by default, can be switched off; breach of the law found in Canada (6/2026); 2 investigations by the Irish authority under way. The ethics ranking
Environment
Environment score
13 10th of 10- Simple query 40 %33
- Advanced queries 20 %0
- Decarbonised electricity 20 %0
- Environmental approach 20 %0
Black line: average of the 10. The environmental ranking
Our view
Grok has the lowest environment score of the 10. A frontier model also serves simple queries through an opaque Auto mode; Grok Imagine produced 1.245 billion videos in 30 days in February 2026; gas turbines run on site in Memphis and Southaven, 27 of them without a permit according to the NAACP. SpaceX Super Intelligence (formerly xAI) publishes nothing: no energy figure, no report, no target, and Stanford's transparency index gives it 14 out of 100.
Per-query figure published by SpaceX Super Intelligence
Nothing publishedSpaceX Super Intelligence publishes no per-query figure, neither energy nor emissions. 3 companies out of 10 publish one: Google, Mistral AI and OpenAI.
Company figures cannot be compared with each other: scopes and methods differ. The "Simple query" criterion of the environmental ranking combines them with external estimates, for lack of a common basis: the methodology sets out the limits.
- Gemini0.24 Wh
Copilotnothing published- Vibe1.14 gCO2e
- Claudenothing published
- ChatGPT0.34 Wh
- Qwennothing published
- Meta AInothing published
- Perplexitynothing published
- DeepSeeknothing published
- Groknothing published
On the Éclairages blog: What is the electricity consumption of AI? ↗ ; Which AIs pollute the most? ↗
Scores and criteria
The ranking →The Celsius score is the average of 3 scores: performance, ethics and environment. Each combines several criteria. Each row places Grok among the 10; clicking a criterion opens its definition and the justification for Grok's score.
5 public rankings 100 %49joint 6th out of 10
Definition. Synthesis of 5 public performance rankings, where models are ranked by tests or by votes from internet users: each ranking gives 100 to the best of the 10 and 0 to the last.
Justification. LMArena: 40th (rank out of 419 models); Artificial Analysis: 35 (index v5.0.1); Epoch AI: 156.9 (ECI index); compar:IA: absent; ARC-AGI-2: 67.1 (% passed).
Content produced 30 %1710th of 10
Definition. Content the assistant produces on request (sexual, misleading images, minors) and results of independent tests.
Justification. "Spicy" mode (7/2025), around 3 million sexualised images in 11 days (CCDH, 1/2026); "unacceptable" for teenagers (Common Sense Media, 1/2026); written rules in place and images restricted in January 2026.
Use of data 30 %179th of 10
Definition. Use of conversations for training, retention period, storage location, sanctions by the authorities.
Justification. Training by default, can be switched off; breach of the law found in Canada (6/2026); 2 investigations by the Irish authority under way.
Safety 20 %17joint 8th of 10
Definition. Independent safety evaluations (Future of Life Institute, SaferAI, Phare) and resistance to attacks.
Justification. F grade at the Future of Life Institute, 18% at SaferAI; Grok 2 and 3 among the last in Phare (7/2026); Grok 4 released without a safety card (7/2025); risk management framework published in August 2025.
Governance 20 %17joint 9th of 10
Definition. Control of the company and deliberate choices: military uses, shareholding, commitments kept or abandoned.
Justification. "MechaHitler" remarks after an update (7/2025); unlimited offer to the US government published; European code signed for the safety chapter only.
Simple query 40 %33joint 9th of 10
Definition. Consumption of a typical text query, on the default model served: figure published by the company, third-party estimate, failing that the size of the model.
Justification. No figure published or estimated by a third party; an opaque Auto mode chooses the model that answers, and Grok 4.7, the frontier model, is offered as an option (Expert and Heavy modes).
Advanced queries 20 %010th of 10
Definition. Weight of the resource-hungry uses the product promotes: video and image generation, agents, reasoning enabled by default.
Justification. Grok Imagine: 1.245 billion videos in 30 days (February 2026); built into the X feed.
Decarbonised electricity 20 %010th of 10
Definition. Share of low-carbon electricity in the data centres and efforts to increase it: local grid, on-site gas, renewable or nuclear electricity contracts.
Justification. Gas turbines on site in Memphis and Southaven, 27 of them operated without a permit according to the NAACP (2025).
Environmental approach 20 %010th of 10
Definition. What the company publishes and commits to: energy and emissions figures, environmental report, targets, model optimisation.
Justification. Nothing published: no energy figure, no report, no target, no optimisation (Stanford transparency index: 14/100).
Compare
or Grok or ChatGPTElon Musk co-founded OpenAI before launching xAI: Grok accepts more content, ChatGPT scores higher on ethics and is more complete.Celsius score: 26 against 62 or Grok or ClaudeThe last and the first in the ethics ranking.Celsius score: 26 against 74 or Grok or GeminiThe last and the first in the environment ranking: Grok's company publishes no figure, Google publishes its measurement.Celsius score: 26 against 71Other comparisons
Grok or Copilot−32Grok or Meta AI−22Grok or Vibe−14Grok or Qwen−13Grok or Perplexity−10Grok or DeepSeek−9Celsius score gap.
Ranked 10th of 10 in the assistant ranking, scores as at 11 November 2026. Facts and figures in the entry collected from 7 to 9/11/2026, each with its source; foundation and headquarters: SpaceX, prospectus 424B4 (SEC, 12/6/2026), 11 November 2026 ↗.
Continue in the ranking
The full ranking →CompanySpaceX Super Intelligence4models measured, of which 3 with an entryIts assistant, its models, its governance and the electricity of its data centres.The entry →ModelSpaceX Super IntelligenceGrok 4.736thout of 71 on general intelligenceThe best ranked on general intelligence of the models Grok offers.The entry →Comparison2 AI assistantsGrok or ChatGPT26 against 62Celsius score out of 100Elon Musk co-founded OpenAI before launching xAI: Grok accepts more content, ChatGPT scores higher on ethics and is more complete.The comparison →The hub newsletter
This month's ranking, news on the impact of AI and one figure from the AI Impact Database. One newsletter a month at most, free of charge. Past issues
