Mistral Small 3.1

Last updated: Recently

Model Type

MText

Max Context

128K

Max Output

Parameter Scale

Mistral Small 3.1 24B Instruct is an upgraded variant of Mistral Small 3 2501 , featuring 24 billion parameters with advanced multimodal capabilities. 128,000 token context window.

Input Modality

text, image

Output Modality

text

Inference Speed

137.831 tokens/s

Latest Release Date

3/17/2025

SDK Ecosystem

artificial-analysismanufactureraa-bootstrap

Model Overview Fields

name

Mistral Small 3.1

Release date

3/17/2025

Performance

137.831 tokens/s

model_creator.name

Mistral

AA Evaluation Scores

AA Intelligence Index

14.5

AA Coding Index

13.9

AA Math Index

3.7

MMLU Pro

0.659

GPQA

0.454

HLE

0.048

LiveCodeBench

0.212

SciCode

0.265

Math-500

0.706666666666667

AIME

0.0933333333333333

AIME 2025

0.0366666666666667

IFBench

0.299319727891156

LCR

0.196666666666667

TerminalBench Hard

0.0757575757575758

TAU2

0.251461988304094

Output Speed (tokens/s)

137.831 tok/s

TTFT (s)

0.512s

First Answer Token (s)

0.512s

Mistral

Mistral Small 3.1

$0.10/M$0.23/M

Mistral Small 3.1 24B Instruct is an upgraded variant of Mistral Small 3 2501 , featuring 24 billion parameters with advanced multimodal capabilities. 128,000 token context window.

There are no reviews for this model yet.

artificial-analysismanufactureraa-bootstrap

0.0