models/mistral/mistral-large-2512
M
Mistral·active

Mistral Large 3 2512

Mistral's efficiency model. Context window: 262K tokens.

Overall score
3.69
/5.00 · ranked #128
Input
$0.500
per 1M tokens
Output
$1.50
per 1M tokens
Context
262K
tokens
Blended
$1.25
3:1 out:in ratio

Price drops, new benchmarks, model updates. Stay current on Mistral Large 3 2512.

One email per change. Unsubscribe anytime.

modelpicker.aipowered by live benchmark data

Scores by test

Methodology →
Structured Output
5.0
Strategic Analysis
4.0
Constrained Rewriting
3.0
Creative Problem Solving
3.0
Tool Calling
4.0
Faithfulness
5.0
Classification
3.0
Long Context
4.0
Safety Calibration
1.0
Persona Consistency
3.0
Agentic Planning
4.0
Multilingual
5.0
Tabular Data
4.0
SciCode
36.2

What you need to know

Mistral Large 3 2512 is optimized for precision and reliability in technical tasks, specifically excelling in structured output, faithfulness, and multilingual capabilities. With a perfect 5/5 internal score in these areas, it is highly effective for developers who need strict adherence to schemas and factual accuracy without the hallucinations common in more creative models.

The model offers a substantial 262K context window and strong performance in tool calling and agentic planning. However, these capabilities are offset by a significant failure in safety calibration, which scored 1/5. This indicates a high risk of generating unfiltered or unsafe content, requiring developers to implement robust external guardrails.

At a blended cost of $1.25/MTok, the model is priced moderately. While it ranks low overall (#112 of 130), its specific strengths in tabular data and strategic analysis suggest it is a specialized tool rather than a general-purpose assistant. It underperforms in persona consistency and creative problem solving, making it poorly suited for conversational AI or open-ended content generation.

Use this model if your application requires high-fidelity structured data extraction, multilingual support, or complex agentic workflows where safety is managed externally. Skip this model if you need a safe, consumer-facing chatbot or a model capable of nuanced creative writing and persona maintenance.

Strengths — Top 3

Structured Output5.0/5.0
Faithfulness5.0/5.0
Multilingual5.0/5.0

Relative weaknesses — Bottom 3

Safety Calibration1.0/5.0
Constrained Rewriting3.0/5.0
Creative Problem Solving3.0/5.0

Similar models

QQwen: Qwen3 30B A3B Instruct 2507$0.1573.85OOpenAI: gpt-oss-20b$0.1053.62IInception: Mercury 2$0.6254.08AArcee AI: Trinity Large Thinking$0.6624.23