MiniMax Text-01
MiniMax · 2025
MiniMax's flagship model with 456B MoE parameters and unprecedented 4M token context window.
Quick Facts
Parameters
456B total (MoE)
Context Window
4M tokens
Modalities
text
Open Source
No
Pricing
API from ~$0.25/1M tokens
Released
2025
Developer
MiniMax
About
MiniMax Text-01 is MiniMax's flagship large language model featuring 456 billion total parameters with Mixture-of-Experts architecture and an unprecedented 4 million token context window — the largest of any known AI model by a significant margin. This 4M context window represents a step change in what's possible with AI: you could feed it the entire Harry Potter series, hundreds of research papers, an entire company's internal documentation, or millions of lines of code in a single request and get comprehensive analysis without any loss of context. While competitors are pushing toward 1M or 2M tokens, MiniMax Text-01 achieves 4M, doubling what was previously thought possible. The MoE architecture with 456B total parameters ensures the model has the capacity to meaningfully use this context. Text-01 demonstrates competitive performance on standard benchmarks while offering unique capabilities for long-context applications that no other model can match. Available through MiniMax's API platform at competitive pricing (approximately USD 0.25 per 1M tokens) with enterprise options for dedicated deployment. For researchers processing massive document archives, legal teams reviewing entire case histories, companies analyzing their full codebase in one pass, and any organization dealing with extremely long documents, MiniMax Text-01 offers capabilities that simply don't exist in any other model. The main limitations are primarily Chinese-market focus, smaller ecosystem compared to Western models, and text-only operation without multimodal capabilities. For applications requiring extreme context handling — academic research, legal document analysis, historical archive processing, comprehensive codebase analysis — MiniMax Text-01 is uniquely capable. Its 4M token context window is 2-4x larger than the nearest competitor.
Strengths
- +Largest context window of any model at 4M tokens
- +456B MoE architecture for strong performance
- +Cost-effective pricing for long-context tasks
- +Unique capability for processing massive documents
Weaknesses
- −Limited availability outside China
- −Smaller ecosystem and community
- −Text-only without multimodal capabilities
Best For
Processing extremely long documents and archives
Analyzing entire codebases in one pass
Long-form content generation at massive scale
Enterprise applications needing extreme context
Pricing
API
From ~$0.25/1M tokens
- 4M token context
- Pay-as-you-go
- Long-context optimized
Enterprise
Custom pricing
- Dedicated deployment
- Higher rate limits
- Custom fine-tuning
Technical Specs
Parameters
456B total (MoE)
Context Window
4M tokens
Modalities
text
Languages
Open Source
No
Developer
MiniMax
Released: 2025
Related Models
GPT-4o
OpenAI
OpenAI's flagship multimodal model combining text, vision, and audio in one unified interface.
GPT-5
OpenAI
OpenAI's latest flagship model with enhanced reasoning, larger context, and improved multimodality.
Claude 3.5 Sonnet
Anthropic
Anthropic's balanced model offering strong reasoning, coding, and long-context capabilities.
Claude 4 Opus
Anthropic
Anthropic's most powerful model for complex reasoning, research, and specialized tasks.