88Z AI, 3 November — MOSS Intelligence has released MOSS-Live, a full-duplex real-time interaction system using a front-back-end collaborative architecture. It supports simultaneous listening and speaking, interruptible dialogue, and multimodal capabilities. The system demonstrates low interruption rates and high anti-interference performance in benchmarks, enabling natural real-time voice communication alongside complex reasoning.
At a glance
- Uses a front-back-end collaborative architecture to support simultaneous listening, speaking, and complex reasoning
- Benchmarks show a low interruption rate of 1.44% and an anti-interference rate of 98.17%

The update
MOSS Intelligence has officially launched MOSS-Live, a full-duplex real-time interaction system. The system innovatively adopts a front-back-end collaborative architecture, where the front end focuses on real-time perception and voice communication, while the back end handles complex reasoning tasks. This design allows the AI to listen and speak simultaneously, intelligently judging when to respond or wait, thereby creating a more natural conversational experience. MOSS-Live features robust interaction capabilities, supporting user interruptions with continued speech, proactive interaction, and simultaneous interpretation. Additionally, the system integrates multimodal abilities such as video and screen sharing. In benchmarks like FDB and τ-Voice, MOSS-Live performed excellently, with an interruption rate of only 1.44% and an agreement anti-interference rate of 98.17%, demonstrating superior real-time response and anti-interference performance.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- 模思智能MOSSSource published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.