models/mistral/mistral-small-3-2-24b-instruct
M
Mistral·active

Mistral Small 3.2 24B

Mistral's efficiency model. Context window: 131K tokens.

Overall score
3.23
/5.00 · ranked #142
Input
$0.075
per 1M tokens
Output
$0.200
per 1M tokens
Context
131K
tokens
Blended
$0.169
3:1 out:in ratio

Price drops, new benchmarks, model updates. Stay current on Mistral Small 3.2 24B.

One email per change. Unsubscribe anytime.

modelpicker.aipowered by live benchmark data

Scores by test

Methodology →
Structured Output
4.0
Strategic Analysis
2.0
Constrained Rewriting
4.0
Creative Problem Solving
2.0
Tool Calling
4.0
Faithfulness
4.0
Classification
3.0
Long Context
4.0
Safety Calibration
1.0
Persona Consistency
3.0
Agentic Planning
4.0
Multilingual
4.0
Tabular Data
3.0

What you need to know

Mistral Small 3.2 24B is optimized for operational reliability and structured tasks rather than cognitive depth. Its strongest differentiators are its high performance in tool calling, agentic planning, and structured output, all scoring 4/5. These capabilities, paired with a 256K context window, make it a viable engine for automated workflows and long-document processing where precise execution is more important than nuanced reasoning.

The model struggles with high-level cognition and safety. With scores of 2/5 in strategic analysis and creative problem solving, it is not suited for complex brainstorming or architectural planning. Most notably, its safety calibration is nearly non-existent at 1/5, meaning developers must implement rigorous external guardrails to prevent problematic outputs.

At a blended cost of $0.250/MTok, this model is positioned as a low-cost utility. While it ranks low overall (#126 of 130), its ability to handle constrained rewriting and multilingual tasks efficiently suggests it provides a high return on investment for specific, narrow technical applications rather than general-purpose assistance.

Use this model if you need a cheap, high-context worker for tool-driven agents, structured data extraction, or multilingual rewriting. Skip this model if your application requires complex strategic reasoning, creative synthesis, or built-in safety filters.

Strengths — Top 3

Structured Output4.0/5.0
Constrained Rewriting4.0/5.0
Tool Calling4.0/5.0

Relative weaknesses — Bottom 3

Safety Calibration1.0/5.0
Strategic Analysis2.0/5.0
Creative Problem Solving2.0/5.0

Similar models

QQwen: Qwen3 Coder 30B A3B Instruct$0.2283.23MLlama 3.3 70B Instruct$0.7103.46MCodestral 2508$0.7503.46OGPT-4o$8.133.46