Mistral
Open weights

Mistral Small 4

Fast Mistral production model for chat, extraction, and cost-sensitive agents

Intelligence
11.3
AA Index
Output price
$0.6/1M
In $0.15/1M
Context
256k
tokens
Speed
172
tokens/sec

Performance

Capability profile

Artificial Analysis indices (0–100). Missing dimensions show as 0.

Benchmarks

GPQA Diamond77%
SciCode39%
Humanity's Last Exam10%

Specifications

Provider
Mistral
Family
mistral-small
Released
2026-03-16
Knowledge cutoff
2025-06
Context window
256,000 tokens
Max output
256,000 tokens
Input price
$0.15/1M
Output price
$0.6/1M
Input modalities
text
image
Output modalities
text
Capabilities
Reasoning · Tool calling · File attachments · Open weights

Specs from models.dev · benchmarks from Artificial Analysis.

AI tools built on Mistral

Explore directory tools powered by Mistral models.

Mistral AI logo
Mistral AI
Generic

Mistral AI is a powerhouse in frontier intelligence, offering high-performance, open-weight large language models designed for enterprise-grade customization. It enables organizations to build, fine-tune, and deploy autonomous agents and tailored AI systems with a focus on data privacy and operational control.