Tips & tricks

LLM

MGX offers four large language models, grouped into two tiers — so you can match the model to the task without wasting your credit.

Basic Models

• Deepseek-chat
• GPT-4o-mini

Advanced Models

• GPT-4o
• Claude-Sonnet

Model Selection Strategy

Not all models burn through quota the same way. Advanced models are much more powerful — and more expensive (in terms of usage). Here's a smart strategy:

• Use basic models for everyday work and frequent tasks.
• Switch to advanced models only when you need depth, precision, or creativity at scale.

Scenarios for Basic Models

Prioritize Deepseek-chat or GPT-4o-mini in these scenarios:

Basic Interaction

• Daily conversations/simple inquiries.
• General knowledge Q&A.
• High-frequency repetitive dialogues.

Content Generation

• Basic text generation.
• Text polishing/simple writing.
• Preliminary creative ideation.

Technical Assistance

• Simple code support.
• Incremental code modifications.
• Basic image analysis (GPT-4o-mini exclusive).

Data Processing

• Basic data analytics.
• Simple modeling tasks.

Scenarios for Advanced Models

Prioritize GPT-4o or Claude-Sonnet in these scenarios:

Professional Development

• High-quality code development/debugging
• Complex system architecture design

Academic Research

• Professional paper writing
• In-depth domain-specific consultations

Intelligent Analysis

• Multidimensional data analytics.
• Complex mathematical modeling.
• Advanced image analysis.

High-Level Cognitive Tasks

• Multi-turn complex dialogues.
• Logical reasoning challenges.
• Innovative solution design.

Professional Services

• Precision multilingual translation.
• Industry-specific deep-dive consultations.