> ## Content Index
> Fetch the complete content index at: https://www.testingcatalog.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Mistral AI debuts Magistral models focused on advanced reasoning
- URL: https://www.testingcatalog.com/mistral-ai-debuts-magistral-models-focused-on-advanced-reasoning/
- Published: 2025-06-10T14:00:41.000Z
- Updated: 2025-06-10T14:58:40.000Z
- Author: Alexey Shabanov
- Tags: Mistral, Latest AI News

Mistral AI has introduced a new family of models under the “Magistral” name, officially unveiled on June 10\. These models focus on strong reasoning capabilities and come in two versions: Magistral Medium and Magistral Small. Magistral Medium is a proprietary model aimed at delivering state-of-the-art performance with reduced size and cost compared to competitors. Magistral Small will be released as an open-source version for the broader community, with final naming potentially subject to change.

> BREAKING 🚨: Mistral AI released 2 new models of the Magistral family with reasoning capabilities.   
>  
> What's new:  
> \- Magistral Small: An open-weight version  
> \- Magistral Medium: An enterprise version in preview  
> \- A new Le Chat Think button  
> \- Flash Reasoning 10x the speed [https://t.co/DpV4YmuX7P](https://t.co/DpV4YmuX7P?ref=testingcatalog.com) [pic.twitter.com/ENLiq1zTPS](https://t.co/ENLiq1zTPS?ref=testingcatalog.com)
> 
> — TestingCatalog News 🗞 (@testingcatalog) [June 10, 2025](https://twitter.com/testingcatalog/status/1932441849022795842?ref%5Fsrc=twsrc%5Etfw&ref=testingcatalog.com)

Magistral Medium has already appeared in internal references (magistral-medium-2506) as the default model for reasoning tasks. The public Le Chat interface now includes a new “Think” button, hinting at its integration and support for transparent, step-by-step reasoning. Behind the scenes, traces in the configuration suggest connections to accelerated answer modes, anonymous inference, and speculative decoding, highlighting performance and privacy as design goals.

![Mistral](https://storage.ghost.io/c/2a/1b/2a1b1782-8506-4d7d-bf53-ad3fb52e2a0f/content/images/2025/06/Le-Chat-Mistral-AI-06-10-2025_03_42_PM.jpg)

Mistral, known for its open model strategy and efficient transformers like Mistral 7B and Mixtral, is now making a direct move into high-reasoning AI. Magistral Medium positions itself against OpenAI’s o3 model and Anthropic’s Claude 4 Opus, both of which have recently emphasized complex reasoning and tool integration. With its dual-track release strategy, Mistral is addressing both enterprise-grade applications and open research needs, keeping flexibility at the center of its roadmap.

![Mistral](https://storage.ghost.io/c/2a/1b/2a1b1782-8506-4d7d-bf53-ad3fb52e2a0f/content/images/2025/06/reasoning_graph.png)

Image: Mistral AI

![Image: Mistral AI - Evals](https://storage.ghost.io/c/2a/1b/2a1b1782-8506-4d7d-bf53-ad3fb52e2a0f/content/images/2025/06/plot-magistral.png)

Image: Mistral AI - Evals

The models are being gradually deployed across Le Chat and select APIs, with broader access expected to follow the announcement.

[Source](https://mistral.ai/news/magistral?ref=testingcatalog.com)