Mistral AI: Faster, Traceable Reasoning with Magistral Models
- Mistral AI, a leading European artificial intelligence company, has announced the release of its Magistral reasoning model.
- The Magistral model is engineered to excel in domain-specific tasks within fields like law, finance, and engineering.Key features include obvious thought processes, allowing users to see how the...
- In benchmark tests, Mistral compared its models against DeepSeek R1, highlighting its performance on the AIME2024, which assesses reasoning capabilities.Magistral Medium scored 73.6% (90% with majority voting @64),...
Mistral AI leaps forward with its Magistral reasoning model, offering faster and more traceable AI solutions for enterprises. This innovative model, available in small (open-source) and Medium (enterprise) versions, excels in critical fields like law, finance, and engineering, providing clear insight into its decision-making processes. Magistral’s multilingual reasoning capabilities further enhance its utility. The design prioritizes clear AI, a significant advantage for compliance. News Directory 3 readers will discover a game-changer with traceable reasoning and rapid response times. Magistral Small is available now via Hugging face and the Medium version via Le Chat and API; expect availability on major cloud platforms soon. Discover what’s next for AI reasoning.
Mistral AI Unveils Magistral Reasoning Model for Enterprises
Updated June 11, 2025
Mistral AI, a leading European artificial intelligence company, has announced the release of its Magistral reasoning model. The new model is available in two versions: magistral Small, an open-source option with 24B parameters, and Magistral Medium, designed for enterprise applications.
The Magistral model is engineered to excel in domain-specific tasks within fields like law, finance, and engineering.Key features include obvious thought processes, allowing users to see how the model arrives at its conclusions.It also supports multilingual reasoning, improving performance when answering queries in languages other than English. This makes Magistral a powerful tool for organizations seeking reliable and auditable AI solutions.
In benchmark tests, Mistral compared its models against DeepSeek R1, highlighting its performance on the AIME2024, which assesses reasoning capabilities.Magistral Medium scored 73.6% (90% with majority voting @64), while the Small version scored 70.7% (83.3% respectively), demonstrating strong reasoning capabilities.
Like DeepSeek R1, Magistral reveals its reasoning process before providing an answer. This feature is available across multiple languages, including English, French, Spanish, German, Italian, arabic, Russian, and Simplified Chinese. A “Flash Answers” mode also speeds up the reasoning process,delivering responses up to 10 times faster than competitors.
Magistral Small is available under the Apache 2.0 license via Hugging Face. Magistral Medium can be accessed through Mistral’s Le chat and via API on La Plateforme. It will soon be available on Amazon SageMaker, IBM WatsonX, Azure AI, and Google Cloud Marketplace.
What’s next
Organizations in legal research,financial forecasting,and other regulated industries can leverage Magistral for more reliable and transparent AI-driven insights. The model’s traceable reasoning makes it especially suitable for compliance and auditability in regulated sectors. Interested users can explore Magistral on Mistral’s Le Chat platform.