Magistral AI: French artificial intelligence that can reason logically and is 10x faster than the competition

Mistral AI, the French leader in artificial intelligence, introduced a new family of language models called Magistral on June 10. This artificial intelligence is distinguished by its logical reasoning and is designed for more complex tasks that require multi-step solutions. Its capabilities put it among leading models such as o3-Pro from OpenAI or Gemini 2.5 Pro from Google.

The first “reasoning” models

While traditional artificial intelligence models simply serve up answers to your questions, so-called “reasoning” models let you look under the hood of the thought process. In practice, this means that artificial intelligence not only provides a solution to your problem, but also transparently shows you the individual steps taken by the AI and how it reached its conclusion.

Mistral AI named this new artificial intelligence Magistral, and it is represented by two models—Magistral Small and Magistral Medium.

Magistral Small and Magistral Medium

Both models are highly reliable and consistent in the quality of their outputs, and they excel in specific areas such as mathematics and physics.

The Magistral Small model is the basic open-source model and can be found on the Hugging Face AI developer platform, where it is freely available under the Apache 2.0 license.

Magistral Medium, by contrast, is a more capable and advanced commercial version of the model, and you can find it in Le Chat, Mistral AI’s chatbot.

Above-average speed and multilingual capabilities

However, Mistral AI has not focused solely on logical reasoning and structured calculations. This new reasoning AI has two other major strengths—speed and knowledge of a large number of languages.

According to the company, the new Magistral model delivers its answers 10x faster than the competition. You can also choose the response speed yourself by selecting either Think mode or Flash Answers.

Among other things, Magistral has been trained in several languages. It will impress you not only with its multi-step computational logic, but also with the naturalness and maximum clarity of its answers. Magistral has an excellent command of not only French and English, but also a deep knowledge of Arabic, Italian, Russian, and Simplified Chinese.

Transparent reasoning

So how does the new model focused on logical reasoning work? It lets you examine its chain of reasoning. While traditional artificial intelligence thinks behind the scenes, this AI allows users to delve into its logical reasoning and the chaining of its thoughts.

This multi-step logic mirrors human thinking, producing clearer and more reliable conclusions. The entire decision-making process and its transparency also allow users to continue working with the individual steps of the process and better understand the result.

Uses of Mistral AI Magistral

Thanks to its structured calculations, transparent reasoning, accuracy, reliability, and natural language, Magistral AI is suitable for a wide range of tasks and fields.

According to the information available so far, it is suitable for education, customer support, business management, data engineering, legal and financial analysis, research and strategic planning, risk assessment, and data-driven decision-making.

In addition, Magistral is a capable creative partner that produces highly creative texts. It can therefore handle a truly broad range of tasks, from mathematics to marketing.

Better than the competition?

Although Magistral achieves above-average results and outperforms some competitors, it still lags slightly behind the leaders of the AI world. According to benchmark tests available on Mistral AI’s website, it outperformed DeepSeek with excellent results in the AIME (mathematics), GPQA Diamond (natural science questions), and LiveCodeBench (coding) tests.

However, when compared with the o3-Pro (OpenAI), Claude Opus 4 (Anthropic), or Gemini 2.5 Pro (Google) models, we can see that it achieved worse results in the same tests. Therefore, although this new AI model is highly capable, it is still not a leader in its field.

Is Magistral AI worth your attention?

The new Magistral language models show that AI no longer has to merely predict words, but can actually think. The test results also confirm that this model undoubtedly excels in both logical reasoning and speed.

However, Magistral’s strength lies not only in its capabilities, but also in the company behind it. Mistral AI has positioned itself as an ambitious developer that understands the needs of businesses and delivers practical AI assistants through its Le Chat Enterprise tool. These assistants can work with enterprise systems or services such as Gmail, significantly simplifying everyday tasks.

If you are looking for powerful, transparent, and customizable artificial intelligence—whether for business deployment or simply for personal use—Magistral AI is definitely worth your attention.

Additional sources: TechCrunch, VentureBeat