Technology1 min read

Mistral Large 4

By · Published by Everything Blog

In short

The Mistral Large 4 is a cutting-edge model designed for general-purpose tasks, featuring a granular Mixture-of-Experts architecture with 52 billion parameters and a total of 1.05 trillion parameters. It excels in vision tasks with a 1.6 billion vision encoder, and its public preview release on October 6, 2026, comes with a price of $1.36 and a performance speed of 1 million tokens per second. This model is a significant advancement in AI technology.

Key points

  • The Mistral Large 4 is a cutting-edge model designed for general-purpose tasks.: The Mistral Large 4 is a cutting-edge model designed for general-purpose tasks.
  • It features a granular Mixture-of-Experts architecture with 52 billion parameters and a t…: It features a granular Mixture-of-Experts architecture with 52 billion parameters and a total of 1.05 trillion parameters.
  • It excels in vision tasks with a 1.6 billion vision encoder.: It excels in vision tasks with a 1.6 billion vision encoder.

Sources checked by our analysis pipeline: 1

October 6, 2026

Public Previewv26.10

Mistral Large 4

Mistral Large 4 is a state-of-the-art, open-weight, general-purpose multimodal model with a granular Mixture-of-Experts architecture. It features 52B active parameters and 1.05T total parameters, and a 1.6B vision encoder.

Context

1M

Price

$1.36

Input/M Tokens

$0.14

Cached input/M Tokens

$4.18

Output/M Tokens

Speed

Performance

Context

1M

Price

$1.36

/M Tokens

$0.14

/M Tokens

$4.18

/M Tokens

Other Models

Original source: docs.mistral.ai

Technology