Vibe
Identity card
The company
- Name
- Mistral AI ↗
- Founded
- 2023
- Head office
- Paris, France
- Chief executive
- Arthur Mensch
The product
- Flagship model
- Mistral Large
- Subscription
- €17.99 per month · Pro
- API
- 0.50 $ / 1.50 $ input / output, per million tokens
The data
- Law
- France and European Union
- Storage
- European Union, by default
Usage
- Web traffic
- not measured
- Developers
- 2.2 %
Key points
Vibe, formerly Le Chat from Mistral AI, is the European alternative: a French company, data hosted in the Union and the lowest API output price of the 9 assistants that have one. Its models remain behind the best, and Mistral AI publishes very little on safety.
- Data in Europe
- French public bodies and companies
- Small budgets
- The most complex tasks
- Documented safety
2.2 % of tokens processed on OpenRouter, the developers' platform, week of 26/10/2026 (6th of 7 assistants present).
Source: Similarweb, data from August 2025 to September 2026, relayed by the press.
Yes if your data must stay in Europe or if you are looking for the best price. Look elsewhere for the tasks that demand the most reasoning or code (Claude, Gemini).
Strengths
- Data hosted in EuropeData is hosted in the European Union by default, according to Mistral AI, and the company has signed the European code of practice on AI in full.Source: Mistral AI, help centre (hosting), 7 November 2026 ↗ ; European Commission, signatories of the AI code, 7 October 2026 ↗
- The only one to publish a life cycle assessment of its modelIn July 2025, with Carbone 4 and ADEME, Mistral AI published the life cycle assessment of its model: 1.14 gCO2e for a response of 400 tokens, about half a page.Source: Mistral AI, 22 July 2025 ↗
- The French state's choiceMistral's models power the AI Assistant rolled out to State civil servants on 16/6/2026, on SecNumCloud-qualified servers.Source: Le Monde Informatique, 16 June 2026 ↗
Weak points
- Behind the best modelsIts best model is 113th of 419 in the LMArena ranking and passes none of the ARC-AGI-2 tasks.Source: LMArena, text ranking, 28 October 2026 ↗ ; ARC Prize, ARC-AGI-2, 28 October 2026 ↗
- Lowest safety gradeRated for the 1st time by the Future of Life Institute in summer 2026, Mistral AI receives an F, the lowest grade, like xAI and DeepSeek, with the lowest score of the 3 and no published safety framework.Source: Future of Life Institute, AI Safety Index, July 2026 ↗
- Russian narratives repeatedNewsGuard found in April 2026 that Le Chat repeated false Russian claims 50% of the time in English and 56.6% of the time in French.Source: Euronews, 16 June 2026 ↗
The company Mistral AI
The Mistral AI entry →Company history
Mistral AI was founded in Paris in April 2023 by 3 researchers who had worked at Google DeepMind and Meta: Arthur Mensch, Guillaume Lample and Timothée Lacroix. Their idea: a European AI, with partly open models, not dependent on the American giants.
The story continues
Le Chat is released in February 2024. The company raises funds quickly: €1.7 billion in September 2025 with the Dutch company ASML, then €3 billion in September 2026 with the South Korean company Samsung.
In May 2026, Le Chat becomes Vibe, an assistant geared towards work and code.
Sources: Wikipedia, 7 November 2026 ↗ ; Mistral AI, 9 September 2025 ↗ ; France Transactions, 8 September 2026 ↗ ; heise online, 29 May 2026 ↗
Leadership
- Chief executive officer and co-founder
- Arthur Mensch ↗
- Chief Science Officer and co-founder
- Guillaume Lample ↗
- Chief Technology Officer and co-founder
- Timothée Lacroix ↗
Key figures
- Valuation, September 2026
- €21.3 billionFrance Transactions, 8 September 2026 ↗
- Latest funding round, September 2026
- €3 billionFrance Transactions, 8 September 2026 ↗
- Revenue target, 2026
- €1 billionFrance Transactions, 8 September 2026 ↗
Vibe plans
- FreeFree0 €0 $
- ProStandard17.99 €14.99 $
- TeamTeam, per seat29.99 €24.99 $
- EnterpriseCompanyon quote
Mistral AI pricing pages: including VAT in France, excluding tax in the United States; up to date as at 9 November 2026.
Content of the 4 plans, prices in France and the United States
Prices in France and the United States. In the United States, prices are shown in dollars and excluding tax; in France, they include 20% VAT. So $20 does not equal €20: the French price is often slightly higher.
- VibeMistral Large1.50 $0.50 $the lowest output price of the 9
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 $
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.
Vibe models
- Company
- Mistral AI
- Models offered
- Mistral Medium 3.5by default, all; 70th of 71 for general intelligence (official page ↗)Mistral Medium 3.5paid plan, Pro, Team, Enterprise (Work mode and remote code agents); 70th of 71 for general intelligence (official page ↗)
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 Mistral AI handles conversations
- Training models on conversationsVibe: yes by default, can be turned off. Vibe Enterprise: no by default. In teams, the administrator can turn off training for the whole organisation.Source: Mistral AI, help centre, 4 November 2026 ↗
- Opting out of trainingTurn off "Allow your interactions to be used to train our models" in the settings, or "Enable data sharing" in the mobile app. A thumbs up or down with a comment does, however, authorise use of the rated conversation.Source: Mistral AI, help centre (opting out of training), 7 November 2026 ↗
- Retention periodFor the duration of processing, with no stated period for conversations; some account data is kept longer, such as identity, 5 years after the end of the contract.Source: Mistral AI, help centre (retention), 7 November 2026 ↗
- Applicable lawFrench and European law (GDPR): Mistral AI is a French company; data is hosted in the Union by default and may, depending on the function used, be transferred outside the Union for a time, under the clauses provided for by the GDPR.Source: Mistral AI, help centre (hosting), 7 November 2026 ↗
Score on this criterion: 58 out of 100, joint 1st out of 10. Data hosted in the European Union by default; training by default, can be turned off; retention period not published. The ethics ranking
Environment
Environment score
72 2nd out of 10- Simple query 40 %67
- Advanced queries 20 %75
- Decarbonised electricity 20 %75
- Environmental approach 20 %75
Black line: average of the 10. The environmental ranking
Our view
Vibe serves a fast mode by default, without video, on servers in Europe (210 gCO2e/kWh, against 384 in the United States). Mistral AI is the only company to have published a life cycle assessment of its model: 1.14 gCO2e for a half-page response. This assessment dates from 2025 and has not been updated, and the company does not publish an environmental report.
Per-query figure published by Mistral AI
1.14 gCO2eper query, according to Mistral AI: life cycle assessment published in July 2025, half-page response.
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 Vibe among the 10; clicking a criterion opens its definition and the justification for Vibe's score.
5 public rankings 100 %710th 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: 113th (rank among 419 models); Artificial Analysis: 26 (index v5.0.1); Epoch AI: 150.6 (ECI index); compar:IA: 29th (rank among 121 models); ARC-AGI-2: 0 (% passed).
Content produced 30 %33joint 7th of 10
Definition. Content the assistant produces on request (sexual, misleading images, minors) and results of independent tests.
Justification. The policy does not prohibit erotica between adults; age 13 without verification; Russian propaganda repeated 56.6% of the time in French (NewsGuard, 2026).
Use of data 30 %58joint 1st out of 10
Definition. Use of conversations for training, retention period, storage location, sanctions by the authorities.
Justification. Data hosted in the European Union by default; training by default, can be turned off; retention period not published.
Safety 20 %17joint 8th of 10
Definition. Independent safety evaluations (Future of Life Institute, SaferAI, Phare) and resistance to attacks.
Justification. Last of the 9 companies rated by the Future of Life Institute (F, summer 2026); no published safety framework; Mistral Large 3, worst resistance to jailbreaks in Phare (24.9%, 7/2026).
Governance 20 %503rd out of 10
Definition. Control of the company and deliberate choices: military uses, shareholding, commitments kept or abandoned.
Justification. French company, European code signed in full; framework agreement with the Ministry of the Armed Forces with no published limit (1/2026).
Simple query 40 %672nd out 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. Fast mode by default, reasoning on demand; 1.14 g CO₂e per half-page response (LCA, 2025).
Advanced queries 20 %751st out of 10
Definition. Weight of the resource-hungry uses the product promotes: video and image generation, agents, reasoning enabled by default.
Justification. No video; free images (Flux); the Work agent reasons on its own.
Decarbonised electricity 20 %751st out 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. Data hosted in the European Union by default (210 g CO₂e/kWh); data centre in Essonne announced.
Environmental approach 20 %752nd out of 10
Definition. What the company publishes and commits to: energy and emissions figures, environmental report, targets, model optimisation.
Justification. Only life cycle assessment published by an AI company (2025), never updated; no corporate report.
Compare
or Vibe or ClaudeVibe keeps data in Europe for less; Claude performs far better.Celsius score: 40 against 74 or Vibe or ChatGPTVibe keeps data in Europe; ChatGPT is far more widely used and better rated on performance.Celsius score: 40 against 62 or Vibe or GeminiThe top 2 of the environmental ranking: Gemini publishes its energy use with its method, Vibe the life cycle assessment of its model.Celsius score: 40 against 71Other comparisons
Vibe or Copilot−18Vibe or Meta AI−8Vibe or Qwen+1Vibe or Perplexity+4Vibe or DeepSeek+5Vibe or Grok+14Celsius score gap.
Ranked 6th 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: Registre national des entreprises, Mistral AI (SIREN 952,418,325), 9 November 2026 ↗.
Continue in the ranking
The full ranking →CompanyMistral AI3models measured, of which 1 with an entryIts assistant, its models, its governance and the electricity of its data centres.The entry →Comparison2 AI assistantsVibe or Claude40 against 74Celsius score out of 100Vibe keeps data in Europe for less; Claude performs far better.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
