9B
40,960 tokens
Apache 2.0
Official open-weight omnimodal MiniCPM release built end-to-end from SigLip2, Whisper-medium, CosyVoice2, and Qwen3-8B into a single 9B-parameter model. Release materials position MiniCPM-o 4.5 as the strongest model in the MiniCPM-o line, supporting text, image, video, and audio inputs with text and speech outputs, plus both instruct and thinking modes in one checkpoint. The model card highlights full-duplex multimodal live streaming so video/audio input and text/speech output can run concurrently without blocking, along with proactive interaction, bilingual real-time speech conversation in English and Chinese, configurable voices, and zero-shot voice cloning. Official benchmarks claim 77.6 average on an 8-benchmark OpenCompass vision evaluation, text average 82.1 across IFEval-PLS, BBH, CMMLU, MMLU, HumanEval, MBPP, Math500, and GSM8K, strong OmniDocBench document parsing, MMHal-Bench trustworthiness near Gemini 2.5 Flash, and broad multilingual support across 30+ languages. The released ecosystem also emphasizes practical deployment through PyTorch, vLLM, SGLang, llama.cpp, Ollama, AWQ and GGUF quantizations, WebRTC demo tooling, and FlagOS multi-chip adaptations.