為不同工作負載選擇合適的 MiMo 模型
我們先把模型落地頁架構搭起來，後續你可以持續往每個模型頁補上場景、示例、對比與 FAQ，而不需要再重做路由。
已準備好的模型頁骨架
下面每個卡片都對應獨立、可持續擴充的 SEO 路由，之後直接補內容即可。
MiMo-V2.5-Pro
可用Flagship reasoning for coding and agent workflows
The primary model for complex agent execution, coding, long-context reasoning, and tool-heavy workflows.
- 上下文視窗
- 1M
- 輸出視窗
- 128K
能力標籤
Text generationDeep reasoningStreamingFunction callingStructured outputWeb search
MiMo-V2.5
可用Full-modal understanding with 1M context
Built for applications that need to understand text, images, video, and audio in one model.
- 上下文視窗
- 1M
- 輸出視窗
- 128K
能力標籤
Full-modal understandingDeep reasoningStreamingFunction callingStructured outputWeb search
MiMo-V2.5-TTS
可用Expressive text-to-speech with built-in voices
Generates natural speech from assistant messages, with style control through instructions and audio tags.
- 上下文視窗
- 8K
- 輸出視窗
- 8K
能力標籤
Text-to-speechAudio outputSingingStyle control
MiMo-V2.5-TTS-VoiceClone
可用Voice cloning from audio samples
Replicates a target voice from an audio sample and uses it for speech synthesis.
- 上下文視窗
- 8K
- 輸出視窗
- 8K
能力標籤
Text-to-speechVoice cloningAudio output
MiMo-V2.5-TTS-VoiceDesign
可用Custom voice design from text descriptions
Creates a voice from a text description, then synthesizes speech in that custom voice.
- 上下文視窗
- 8K
- 輸出視窗
- 8K
能力標籤
Text-to-speechVoice designAudio output
MiMo-V2-Pro
可用Reasoning and production-grade text generation
Built for agent workflows, structured output, and long-context reasoning tasks.
- 上下文視窗
- 1M
- 輸出視窗
- 128K
能力標籤
Text generationDeep reasoningStreamingFunction callingStructured outputWeb search
MiMo-V2-Omni
可用Multimodal understanding for image, audio, and richer inputs
Designed for teams building assistants and applications that need multimodal perception.
- 上下文視窗
- 256K
- 輸出視窗
- 128K
能力標籤
Multimodal understandingDeep reasoningStreamingFunction callingWeb search