MiniMax M3

minimax-m3

MiniMax's flagship model — code, reasoning and analysis with a 1 million token context.

代码推理分析

技术规格

API 标识符	`minimax-m3`
订阅计划	Pro
最大上下文	1 000 000 tokens
视觉（图像）	否
原生网络访问	否
图像生成	否
输入成本	$1.50 / M tokens
输出成本	$6.00 / M tokens

概述

MiniMax M3 is MiniMax's most advanced model, succeeding M2.7. It combines excellent code, reasoning and analysis capabilities with the signature context window of the MiniMax family: 1 million tokens. Mingo reserves it for the Pro plan for projects requiring both high quality and the ability to process huge amounts of information in a single request.

优势

MiniMax flagship1M token contextCode and reasoningIn-depth analysis

理想使用场景

Whole codebase analysis

Analyze and understand entire software projects in a single request.

Massive documentation synthesis

Summarize and cross-reference very large sets of documents.

Reasoning over large volumes

Solve complex problems requiring a lot of context.

Large-scale data analysis

Process and analyze very extensive datasets.

定价

输入 Token

$1.50

每百万 Token

输出 Token

$6.00

每百万 Token

常见问题

Does MiniMax M3 replace MiniMax M2.7?

MiniMax M3 is a new, more capable generation reserved for the Pro plan, while M2.7 and its High Speed variant remain available on Basic.

Does MiniMax M3 really have 1M tokens of context?

Yes, like the entire MiniMax family, it has a 1 million token context window.

When does Mingo choose MiniMax M3?

For the most demanding code, reasoning or analysis tasks requiring both high quality and a very large context.

Use MiniMax M3 with Mingo

Mingo routes your most demanding projects requiring a large context to MiniMax M3 on the Pro plan.

免費試用 Mingo