Description
The reasoning post-trained version of Ministral 3 3B, combining a 3.4B language model with a 0.4B vision encoder, optimized for complex reasoning tasks.
Stats
633 Downloads
2 stars
Capabilities
Minimum system memory
Tags
Last updated
Updated 2 days agobyREADME
Ministral 3 3B Reasoning by mistralai
The reasoning post-trained version of Ministral 3 3B, combining a 3.4B language model with a 0.4B vision encoder, optimized for complex reasoning tasks.
Supports context length of 256k tokens.
Excels at complex, multi-step reasoning and dynamic problem-solving, making it ideal for math, coding, and STEM-related use cases.
Vision-enabled for image analysis and multimodal reasoning tasks.
Multilingual support across dozens of languages including English, French, Spanish, German, Italian, Portuguese, Dutch, Chinese, Japanese, Korean, Arabic.
Native function calling and JSON output generation with best-in-class agentic capabilities.
Edge-optimized for deployment on a wide range of hardware including local devices.
Apache 2.0 License
Parameters
Custom configuration options included with this model
Sources
The underlying model files this model uses
Based on