MiniMax-01 FAQs

Question 1

What kind of attention mechanisms are used in MiniMax-01?

Accepted Answer

MiniMax-01 employs a hybrid attention architecture, including Lightning Attention, Softmax Attention, and Mixture-of-Experts (MoE), optimized for performance and long-context capabilities.

Question 2

What is MiniMax-VL-01, and what are its capabilities?

Accepted Answer

MiniMax-VL-01 is a vision-language model built on MiniMax-Text-01, utilizing a ViT-MLP-LLM framework. It excels in multimodal tasks thanks to its dynamic resolution mechanism for image processing.

Question 3

What are the key features of MiniMax-Text-01?

Accepted Answer

MiniMax-Text-01 boasts 456 billion parameters, a hybrid attention architecture (Lightning, Softmax, MoE), and a training context length of 1 million tokens, enabling high performance on text-based tasks.

Question 4

What is MiniMax-01?

Accepted Answer

MiniMax-01 provides access to large language and vision-language models, including MiniMax-Text-01 and MiniMax-VL-01, designed for various natural language and multimodal tasks.

Question 5

Where can I find the license information for the MiniMax-01 models and code?

Accepted Answer

The model license agreement can be found at [Model License Link], and the code license (MIT) is available at [Code License Link] on the GitHub repository.

MiniMax-01

MiniMax-01

Key Features

Use Cases

Key Features

Use Cases