models/minimax/minimax-01
M
minimax·active

MiniMax: MiniMax-01

minimax's efficiency model. Long-context specialist with 1.0M window.

Overall score
3.08
/5.00 · ranked #144
Input
$0.200
per 1M tokens
Output
$1.10
per 1M tokens
Context
1.0M
tokens
Blended
$0.875
3:1 out:in ratio

Price drops, new benchmarks, model updates. Stay current on MiniMax: MiniMax-01.

One email per change. Unsubscribe anytime.

modelpicker.aipowered by live benchmark data

Scores by test

Methodology →
Structured Output
4.0
Strategic Analysis
3.0
Constrained Rewriting
4.0
Creative Problem Solving
2.0
Tool Calling
Faithfulness
5.0
Classification
3.0
Long Context
4.0
Safety Calibration
1.0
Persona Consistency
2.0
Agentic Planning
3.0
Multilingual
4.0
Tabular Data
2.0

What you need to know

MiniMax-01 is primarily a tool for high-fidelity data extraction and transformation. Its strongest differentiator is a perfect 5/5 faithfulness score, meaning it adheres strictly to provided source material without introducing hallucinations. This reliability is paired with strong performance in structured output and constrained rewriting, making it effective for converting raw text into specific formats.

The model supports a massive 1.0M token context window and is priced aggressively for the mid-tier, with a blended cost of $0.875/MTok. While the input cost is low, the output cost is significantly higher, which is a critical consideration for tasks requiring long-form generation. However, the low cost does not compensate for poor safety calibration and weak creative problem solving.

Performance is inconsistent across general tasks, as evidenced by an overall rank of 127 out of 130 models and a low average internal score of 3.08. It struggles significantly with persona consistency and tabular data, indicating it is not suitable for conversational agents or complex data analysis.

Use this model if you need a low-cost solution for high-accuracy extraction, multilingual rewriting, or processing very large documents where faithfulness is the priority. Skip this model if your application requires strict safety guardrails, creative reasoning, or stable persona maintenance.

Strengths — Top 3

Faithfulness5.0/5.0
Structured Output4.0/5.0
Constrained Rewriting4.0/5.0

Relative weaknesses — Bottom 3

Safety Calibration1.0/5.0
Creative Problem Solving2.0/5.0
Persona Consistency2.0/5.0

Similar models

MMistral Small 3.1 24B$0.5042.77MMistral Small 3.2 24B$0.1693.23IIBM: Granite 4.1 8B$0.0873.54QQwen: Qwen3 Coder 30B A3B Instruct$0.2283.23