Releases·Model Release
inclusionAI releases Realtime-Venus: 9B full-duplex audio-visual omni model on Hugging Face
inclusionAI published Realtime-Venus on Hugging Face: a full-duplex audio-visual interaction system with two checkpoints, including a 9B Omni model adapted from MiniCPM-o 4.5. The system continuously watches and listens, proactively decides when to respond, and handles interruptions on a shared causal timeline. For practitioners it offers streaming speech output, training-free long-video memory, and asynchronous tool delegation via the Realtime-Venus-Harness runtime on GitHub.
145 engagement1 sourceFri, Sep 18, 2026, 03:27 PM
AI Overview
- inclusionAI released Realtime-Venus-Omni, a 9B audio-visual model adapted from MiniCPM-o 4.5, with full-duplex conversation, proactive interaction, semantic interruption handling, and training-free long-video memory
- Realtime-Venus emits in-stream tool requests on a shared causal timeline, executed asynchronously via the Realtime-Venus-Harness runtime on GitHub
- A second checkpoint, Realtime-Venus-Audio, handles audio understanding and audio-driven conversation with text or speech output using bundled Token2wav resources
Related
Developments·Paper
Gander technical report: end-to-end omni model unifies real-time full-duplex interaction and agentic capabilities
A technical report on arXiv introduces Gander, an end-to-end model unifying omni perception, realtime interaction, and agentic capabilities in a single framework. Unlike turn-based paradigms, Gander processes streaming video, speech, and text inputs for full-duplex interaction, supporting interruptions and proactive follow-up questions in both conversations and workflow-oriented agent scenarios.
A technical report on arXiv introduces Gander, an end-to-end model unifying omni perception, realtime interaction, and agentic capabilities in a single framework. Unlike turn-based paradigms, Gander processes streaming video, speech, and text inputs for full-duplex interaction, supporting interruptions and proactive follow-up questions in both conversations and workflow-oriented agent scenarios.
0 engagement1 sourcearxivTue, Sep 8, 2026, 04:22 PM
Developments·Paper
AV-STE: Modular audio-visual front-end restores noisy speech tokens for frozen full-duplex dialogue models
A new arxiv paper addresses robustness failures in full-duplex spoken dialogue systems, whose audio-only perception breaks down under background noise and overlapping speech. It builds on audio-visual dialogue work showing lip-movement cues improve robustness, while avoiding costly multimodal retraining of the dialogue model itself. The approach matters to practitioners building real-time voice agents that must handle noisy, overlapping real-world audio.
A new arxiv paper addresses robustness failures in full-duplex spoken dialogue systems, whose audio-only perception breaks down under background noise and overlapping speech. It builds on audio-visual dialogue work showing lip-movement cues improve robustness, while avoiding costly multimodal retraining of the dialogue model itself. The approach matters to practitioners building real-time voice agents that must handle noisy, overlapping real-world audio.
0 engagement1 sourcearxivTue, Sep 8, 2026, 07:59 AM
Developments·Paper
OmniVChat: Synthesizing, Benchmarking, and Training for Native Audio-Visual Dialogue
An arXiv paper defines OmniVChat, a task where omni models take simultaneous audio and video input directly from a user and return text without separate text questions, captioning, or ASR. The authors built OmniVChat-Studio to synthesize training dialogues, OmniVChat-Bench to evaluate five ability categories, and OmniVChat-RL, a reward design targeting correctness, efficiency, and style. Training Qwen3-Omni-Instruct with OmniVChat-RL improved performance on both synthetic and human-recorded benchmarks.
An arXiv paper defines OmniVChat, a task where omni models take simultaneous audio and video input directly from a user and return text without separate text questions, captioning, or ASR. The authors built OmniVChat-Studio to synthesize training dialogues, OmniVChat-Bench to evaluate five ability categories, and OmniVChat-RL, a reward design targeting correctness, efficiency, and style. Training Qwen3-Omni-Instruct with OmniVChat-RL improved performance on both synthetic and human-recorded benchmarks.
0 engagement1 sourcearxivFri, Sep 18, 2026, 08:15 AM
Releases·Model Release
NVIDIA releases NemotronLabs VoiceChat-11B full-duplex model on Hugging Face
NVIDIA has released the NemotronLabs VoiceChat-11B model on Hugging Face, featuring full-duplex voice interaction. The model enables real-time, bidirectional conversation, allowing users to speak and listen simultaneously. This release advances voice AI capabilities for developers building conversational interfaces.
NVIDIA has released the NemotronLabs VoiceChat-11B model on Hugging Face, featuring full-duplex voice interaction. The model enables real-time, bidirectional conversation, allowing users to speak and listen simultaneously. This release advances voice AI capabilities for developers building conversational interfaces.
249 engagement1 sourceredditMon, Aug 3, 2026, 10:24 PM
Releases·Model Release
🚀 Meet Qwen3.8-Omni-Flash, Qwen's first omni-modal model built around agentic capabilities!
Qwen announced Qwen3.8-Omni-Flash, its first omni-modal model designed around agentic capabilities. The model combines native audio-video understanding, reasoning, and tool use so it can understand content, plan a task, execute with tools, and deliver a result. For practitioners, it signals Qwen's move toward single-model agents spanning perception and action.
Qwen announced Qwen3.8-Omni-Flash, its first omni-modal model designed around agentic capabilities. The model combines native audio-video understanding, reasoning, and tool use so it can understand content, plan a task, execute with tools, and deliver a result. For practitioners, it signals Qwen's move toward single-model agents spanning perception and action.
1.8k engagement2 sourcesx, arxivFri, Sep 18, 2026, 03:16 AM