SpeechGPT-2.0-preview

GPT-4o-level real-time end-to-end spoken-dialogue LLM with <200ms latency, focused on Chinese speech — the flagship continuation of the SpeechGPT lineage from Xipeng Qiu's OpenMOSS group, released pre-SII under Fudan/open-moss.com branding. Models expressive human-like conversation (interruptions, style, emotion) directly in the speech domain rather than cascading ASR→LLM→TTS.