为不同工作负载选择合适的 MiMo 模型
我们先把模型落地页架构搭起来，后续你可以持续往每个模型页里补场景、示例、对比和 FAQ，而不用再重做路由。
已经准备好的模型页骨架
下面每个卡片都对应独立、可持续扩展的 SEO 路由，后面直接补内容就行。
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.
- 上下文窗口
- 1M
- 输出窗口
- 128K
能力标签
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.
- 上下文窗口
- 1M
- 输出窗口
- 128K
能力标签
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.
- 上下文窗口
- 8K
- 输出窗口
- 8K
能力标签
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.
- 上下文窗口
- 8K
- 输出窗口
- 8K
能力标签
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.
- 上下文窗口
- 8K
- 输出窗口
- 8K
能力标签
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.
- 上下文窗口
- 1M
- 输出窗口
- 128K
能力标签
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.
- 上下文窗口
- 256K
- 输出窗口
- 128K
能力标签
Multimodal understandingDeep reasoningStreamingFunction callingWeb search