Mistral logo

Mistral Small 3

Mistral · Released Sep 1, 2025

ProprietarybudgetcodingCompare this model

$0.1

Input price / MTok

#3 of 27

$0.3

Output price / MTok

#3 of 27

128K

Context window

#28 of 28

166

Tokens / sec

#5 of 21

Summary

Mistral Small 3 is a proprietary language model from Mistral, released on Sep 1, 2025.

At $0.1 input and $0.3 output per million tokens, it ranks #3 of 27 for input price, placing it among the most affordable options.

Its context window of 128K tokens ranks #28 of 28 for context depth in this dataset.

It targets budget, coding workloads.

Throughput is measured at 166 tokens per second (ranked #5 of 21 for throughput).

No independently verified benchmark scores are available for this model. Scores are only shown when independently confirmed with a source URL.

Speed source: https://artificialanalysis.ai/models/mistral-small-3