MiniCPM-o 4.5: Towards Real-Time Full-Duplex Omni-Modal Interaction

2026-04-30

AI Summary

Existing multimodal interaction systems typically rely on a turn-taking mechanism that separates perception from response, lacking real-time synchronous processing and proactive environmental intervention capabilities. This work proposes Omni-Flow, a unified streaming framework that, for the first time in an open-source model, enables full-duplex, synchronized input and output across vision, audio, and speech modalities, supporting concurrent perception and response as well as scene-aware proactive behaviors. Built upon an efficient 9B-parameter architecture, Omni-Flow integrates multimodal temporal alignment and edge-optimized inference, achieving visual-language performance comparable to Gemini 2.5 Flash and surpassing Qwen3-Omni-30B-A3B in both multimodal understanding and speech generation, while operating in real time on devices with less than 12GB of memory.

Abstract

Recent progress in multimodal large language models (MLLMs) has brought AI capabilities from static offline data processing to real-time streaming interaction, yet they still remain far from human-level multimodal interaction. The key bottlenecks are no longer modality coverage or latency alone, but the interaction paradigm itself. First, perception and response are still separated into alternating phases, preventing models from incorporating new inputs for timely adjustment during generation. Second, most current models remain reactive, responding only to explicit user requests instead of acting proactively in the evolving multimodal environment. We present MiniCPM-o 4.5, our latest effort towards human-like multimodal interaction, which mitigates these gaps by real-time full-duplex omni-modal interaction. It can see, listen, and speak simultaneously in real-time, while also exhibiting proactive behaviors such as issuing reminders or comments based on its continuous understanding of the live scene. The key technique behind MiniCPM-o 4.5 is Omni-Flow, a unified streaming framework that aligns omni-modal inputs and outputs along a shared temporal axis. This formulation converts conventional turn-based interaction into a full-duplex, time-aligned process, enabling simultaneous perception and response and allowing proactive behavior to arise within the same framework. With a total of 9B parameters, MiniCPM-o 4.5 approaches Gemini 2.5 Flash in vision-language capabilities, delivering state-of-the-art open-source performance at its scale. It also surpasses Qwen3-Omni-30B-A3B in omni-modal understanding and delivers better speech generation, with significantly higher computation efficiency. Driven by its efficient architecture design and inference optimization, the model can perform real-time full-duplex omni-modal interaction on edge devices with less than 12GB RAM cost.

Problem

• multimodal interaction
• full-duplex
• real-time
• proactive behavior
• perception-response alignment

Innovation

• full-duplex interaction
• omni-modal streaming
• Omni-Flow framework
• proactive multimodal behavior
• real-time multimodal LLM