Skip to content

MiniCPM-o 4.5 Model Introduction

This page introduces the MiniCPM-o 4.5 omni-modal architecture, sub-model breakdown, and GGUF file checklist to help you prepare before deployment.

Prerequisite: Recommended to complete ROCm Environment Setup first.


1. What is MiniCPM-o

MiniCPM-o (Omni) is an end-to-end omni-modal model series from OpenBMB. Version 4.5 has approximately 9B total parameters, composed of four backbone modules:

Backbone ModuleBase Model
LLM backboneQwen3-8B
Vision encoderSigLIP2
Audio encoderWhisper-medium
Speech synthesis (TTS)CosyVoice2

The model is built in an end-to-end fashion — the modality encoders/decoders and LLM are densely connected via hidden states, enabling better information flow and control, and facilitating full exploitation of rich multimodal knowledge during training.

Based on this architecture, MiniCPM-o 4.5 integrates the following capabilities:

CapabilityDescription
Text chatSame as a regular LLM, supports multi-turn conversation and system prompts, 30+ languages
Voice inputReal-time audio stream encoding, understands spoken input (bilingual real-time speech in Chinese & English)
Image / Video understandingHigh-resolution images (up to 1.8M pixels) and high-FPS video understanding (up to 10fps)
Voice output (TTS)Natural speech synthesis, supports voice cloning, emotion control, and long speech generation (>1 min)
Full-duplex conversationSimultaneously listens to the microphone and generates voice responses, input and output streams do not block each other
Proactive interactionMonitors input streams at ~1Hz frequency, can initiate unprompted responses based on continuous scene understanding

Compared to a pure text model of the same parameter count, MiniCPM-o 4.5 achieves a full-duplex response latency of approximately 800 ms, enabling near-real-time voice interaction on edge devices.


2. Model Architecture

MiniCPM-o 4.5 consists of multiple independent modules, each loaded as a separate GGUF file by llama.cpp-omni:

MiniCPM-o 4.5 (full omni inference)
├── LLM (Qwen3-8B)             MiniCPM-o-4_5-Q4_K_M.gguf             ~4.9 GB
├── Vision encoder (SigLIP2)    vision/MiniCPM-o-4_5-vision-F16.gguf  ~0.9 GB
├── Audio encoder               audio/MiniCPM-o-4_5-audio-F16.gguf ~1.2 GB
├── TTS language model          tts/MiniCPM-o-4_5-tts-F16.gguf   ~0.5 GB
├── TTS projector               tts/MiniCPM-o-4_5-projector-F16.gguf  ~0.1 GB
└── Token-to-Wav (vocoder)
    ├── token2wav-gguf/encoder.gguf
    ├── token2wav-gguf/flow_matching.gguf
    ├── token2wav-gguf/flow_extra.gguf
    ├── token2wav-gguf/hifigan2.gguf
    └── token2wav-gguf/prompt_cache.gguf    total ~0.7 GB

All 10 GGUF files total approximately 8.3 GB (Q4_K_M quantization).


3. Available Versions and Quantization

This tutorial uses Q4_K_M quantization for the LLM backbone, which balances precision and VRAM usage well for consumer AMD GPUs and APUs.

FileQuantizationPurpose
MiniCPM-o-4_5-Q4_K_M.ggufQ4_K_MLLM backbone (text generation)
vision/MiniCPM-o-4_5-vision-F16.ggufF16Image understanding
audio/MiniCPM-o-4_5-audio-F16.ggufF16Audio input understanding
tts/MiniCPM-o-4_5-tts-F16.ggufF16Speech synthesis (LM part)
tts/MiniCPM-o-4_5-projector-F16.ggufF16Speech synthesis (projection)
token2wav-gguf/*.gguf (5 files)Vocoder (audio waveform generation)

Sub-models (vision, audio, TTS) are currently only available in F16. Even with Q4 quantization on the LLM backbone, total VRAM usage remains ~8.3 GB. Future Q8/Q4 sub-model variants may reduce this further.


4. VRAM Estimation

ModeModules loadedApprox. VRAM
Text onlyLLM~5 GB
Text + imageLLM + vision encoder~6 GB
Text + voice inputLLM + audio encoder~6.5 GB
Full omni (voice in + TTS out)LLM + vision + audio + TTS + Token2Wav~9 GB

Note: Estimates assume a 4096-token context. Using an 8192-token context adds ~1–2 GB for KV cache.

AMD Ryzen AI MAX+ 395 APUs with 64 GB unified memory (all usable as VRAM) can comfortably run all modules with room for large contexts.


5. Conversation Modes

Both llama.cpp-omni and MiniCPM-o-Demo support 4 conversation modes:

ModeInputOutputNotes
Turn-basedAudio file / textText + TTSMost stable; good for initial validation
Half-DuplexLive microphoneText + TTSSpeaks after user finishes; no interruption
Omni (full-duplex)Live mic + cameraReal-time audio streamSimultaneous listen + speak with interruption
Audio DuplexLive microphoneReal-time audio streamSame as Omni but camera not required

CLI deployment (llama-omni-cli) is suited to validating Turn-based mode. The Web Demo supports all 4 modes, enabling real Omni full-duplex via a browser with camera and microphone.


6. Where to Download Models

GGUF files are available from ModelScope (recommended for mainland China, faster download) or Hugging Face:

SourceURL
Hugging FaceOpenBMB/MiniCPM-o-4_5-gguf
ModelScopeOpenBMB/MiniCPM-o-4_5-gguf

References


After reading about the model architecture, proceed to llama.cpp-omni CLI Deployment or Web Demo Full-Duplex Deployment.