Mistral Large 3
LLMLLMMultimodalMultimodal
Open-weight multimodal model from Mistral AI (December 2025), built on a Mixture-of-Experts architecture with 41B active and 675B total parameters, 256k context, Apache 2.0 license.
Technical specification
Context window
Parameters
License
Weights access
Last updated: May 2, 2026
Modalities
Input
Text
Image
Output
Text
Code
Capabilities
7Image understanding★
Vision
Multilingual★
Language
Multimodal understanding★
Multimodality
Function Calling
Planning
Structured output★
Structured gen.
Long context★
Reasoning
OCR★
Vision