Overview
Get answers from minimax-m2 in two modes:- Simple mode (no
response_format) — Returns plain text answers - Structured mode (with
response_format) — Returns JSON matching your schema
Use cases
Simple mode:- Q&A and general knowledge
- Explanations and summaries
- Content generation
- Data extraction from text
- Sentiment analysis
- Email/ticket classification
- Resume/document parsing
- Named entity recognition
- API integration requiring JSON
JSON Schema examples
Theresponse_format parameter accepts standard JSON Schema:
Simple object
Classification with enums
Arrays
Model details
- Model: minimax/minimax-m2 (Novita AI)
- Reasoning: Medium effort
- Max tokens: 65,536
vs Chat Completion
| Feature | /v1/answer | /v1/chat-completion |
|---|---|---|
| Use case | Quick answers, extraction | Conversations, tools |
| Structured output | ✅ Native JSON Schema | ⚠️ Prompt engineering |
| Tool calling | ❌ No | ✅ Yes |
| Multi-turn | ❌ Single turn | ✅ Full history |
Next steps
- Full agentic model: /api-reference/chat/completion
- Available models: /api-reference/models
Body
application/json