リリース2025/12
生成速度0 トークン/秒
初動レイテンシ (TTFT)0.00s
入力料金$0/1M
出力料金$0/1M
ランキング
レビュー
MiMo-V2-Flash is a specialized model for mathematics and reasoning, well-suited for education and research, but weak in programming and general chat. Recommended if you need high mathematical accuracy; avoid if you need coding assistance or open-ended dialogue.
長所
- Exceptional mathematics, ranking #5 on the leaderboard—ideal for complex problems and precise calculations
- Strong logical reasoning, well-suited for analysis and multi-step problem-solving
- Decent AI Agent capabilities, able to execute repetitive automated tasks
短所
- Weak coding performance (ranked #38), making it a poor choice for programming, debugging, or code review
- Poor general chat (ranked #68, lowest), unsuitable for chatbots or general-purpose Q&A
ユースケース
Math problem solving, STEM coursework, and complex formula calculationsData analysis, logical reasoning, and multi-step problem solvingWorkflow automation and AI agents executing repetitive tasks
ガイド・動画
Access MiMo-V2-Flash via OpenRouter (free), GitHub repository, or run it locally using SGLang. The model excels at reasoning, coding, and agent tasks—delivering 150 tokens/second at just $0.1 per 1M input tokens. When using with agentic tools (Claude Code, Cline), disable reasoning mode to optimize speed.