リリース2026/06
生成速度80 トークン/秒
初動レイテンシ (TTFT)1.18s
入力料金$0.3/1M
出力料金$1.2/1M
ランキング
レビュー
MiniMax-M3 is a solid choice for reasoning tasks requiring high precision. While not suited for general-purpose chatbots or automation agents, it excels at complex analysis, logical problem-solving, and consistency checking.
長所
- Exceptional reasoning: ranked #4 AIM, handling complex logic and deep analysis with high precision
- Strong step-by-step reasoning for problems demanding analytical thinking
- Effective at consistency checking and identifying logical flaws
短所
- Limited AI agent capabilities (ranked #63): unsuitable for complex workflow automation
- General chat and coding are not strengths (ranked #17, #22): only average performance
- Context window or processing speed may be limiting for high-volume tasks
ユースケース
Solving logic problems, mathematics, and legal analysis requiring multi-step reasoningAuditing argument quality and detecting logical flaws in documents or proposalsExpert consulting where depth of analysis takes priority over output volume
ガイド・動画
MiniMax-M3 is an advanced 428B-parameter AI model supporting text, image, and video inputs with a 1M-token context window. Access it via code.minimax.io for a coding IDE or through the API at platform.minimax.io/subscribe/token-plan ($20/month). Well-suited for coding, long-document analysis, and automated agent workflows.