Skip to content
@OpenMOSS

OpenMOSS (SII)

OpenMOSS Team is a research group under the Shanghai Innovation Institution (SII), working in close collaboration with Fudan University and MOSI Intelligence.
README.md
OpenMOSS

Shanghai Innovation Institute (SII) · Fudan University · MOSI.AI

Open, collaborative research on Large Language Models and Multimodal Foundation Models.

🌐 Website · 🤗 Hugging Face · ✉️ openmoss@sii.edu.cn

👋 About Us

OpenMOSS is a research group led by Prof. Xipeng Qiu, hosted at the Shanghai Innovation Institute (SII) and working in close collaboration with Fudan University and MOSI.AI. We conduct cutting-edge research across the full LLM stack — from model architecture and training to evaluation, interpretability, and real-world applications — with a strong commitment to open and reproducible science.

🔬 Research Directions

Direction Flagship Repositories
🧠 Foundation LLMs MOSS
👁️ Vision & Video MOSS-VL · MOSS-Video-Preview · MOVA
🔊 Speech & Audio MOSS-TTS · MOSS-TTS-Nano · MOSS-TTSD · MOSS-Audio · MOSS-Speech · MOSS-Audio-Tokenizer
🤖 Embodied AI & Robotics Awesome-WAM · Embodied-Planner-R1 · RoboOmni · FRoM-W1
🔍 Interpretability Llamascopium (formerly Language-Model-SAEs)

✨ Recent Highlights

  • MOSS-TTS-Nano — 0.1B-param multilingual TTS, runs on CPU, 3.2k★
  • MOSS-Audio — Unified audio understanding foundation model (4B / 8B, Instruct & Thinking variants)
  • MOVA — Scalable and synchronized video–audio generation
  • MOSS-VL — Multimodal model series with XRoPE architecture, full training stack open-sourced
  • Awesome-WAM — Curated reading list for World Action Models in embodied AI

See the pinned repositories for quick access, or browse all 50+ repositories.

🤝 Join Us

We welcome researchers, students, and collaborators who share our vision. For PhD/intern openings, research collaborations, or general inquiries, please reach us at openmoss@sii.edu.cn.

The Shanghai Innovation Institute (SII) is dedicated to fostering innovation in education and research in the field of artificial intelligence.

Pinned Loading

  1. MOSS MOSS Public

    An open-source tool-augmented conversational language model from Fudan University

    Python 12.1k 1.1k

  2. MOSS-VL MOSS-VL Public

    MOSS-VL is the core multimodal model series within the OpenMOSS ecosystem, dedicated to visual understanding.

    Python 261 4

  3. MOVA MOVA Public

    MOVA: Towards Scalable and Synchronized Video–Audio Generation

    Python 1k 87

  4. MOSS-TTS-Nano MOSS-TTS-Nano Public

    MOSS-TTS-Nano is an open-source multilingual tiny speech generation model from MOSI.AI and the OpenMOSS team. With only 0.1B parameters, it is designed for realtime speech generation, can run direc…

    Python 3.5k 450

  5. MOSS-Audio MOSS-Audio Public

    MOSS-Audio is an open-source foundation model for unified audio understanding, enabling speech, sound, music, captioning, QA, and reasoning in real-world scenarios.

    Python 574 41

  6. Llamascopium Llamascopium Public

    Performant framework for training, analyzing and visualizing Sparse Autoencoders (SAEs) and their frontier variants.

    Python 221 29

Repositories

Showing 10 of 53 repositories
  • MOVA Public

    MOVA: Towards Scalable and Synchronized Video–Audio Generation

    OpenMOSS/MOVA’s past year of commit activity
    Python 1,045 Apache-2.0 87 30 0 Updated Jun 18, 2026
  • Awesome-WAM Public

    A curated, continuously updated reading list, paper blogs, and resources for World Action Models (WAMs) in embodied AI.

    OpenMOSS/Awesome-WAM’s past year of commit activity
    HTML 851 MIT 21 1 8 Updated Jun 18, 2026
  • MOSS-Video-Preview Public

    A real-time video understanding foundation model with gated cross-attention. Offline & real-time inference.

    OpenMOSS/MOSS-Video-Preview’s past year of commit activity
    Python 151 Apache-2.0 4 0 0 Updated Jun 18, 2026
  • MOSS-VL Public

    MOSS-VL is the core multimodal model series within the OpenMOSS ecosystem, dedicated to visual understanding.

    OpenMOSS/MOSS-VL’s past year of commit activity
    Python 261 Apache-2.0 4 0 0 Updated Jun 18, 2026
  • MOSS-TTS Public

    MOSS‑TTS Family is an open‑source speech and sound generation model family from MOSI.AI and the OpenMOSS team. It is designed for high‑fidelity, high‑expressiveness, and complex real‑world scenarios, covering stable long‑form speech, multi‑speaker dialogue, voice/character design, environmental sound effects, and real‑time streaming TTS.

    OpenMOSS/MOSS-TTS’s past year of commit activity
    Python 3,385 Apache-2.0 292 12 0 Updated Jun 18, 2026
  • claude-codex-handoff Public

    Drop-in async file-based handoff protocol for two AI coding agents (Claude Code + Codex), installed as one shared .handoff/ in your project.

    OpenMOSS/claude-codex-handoff’s past year of commit activity
    Python 21 MIT 0 0 0 Updated Jun 17, 2026
  • MOSS-Audio-Tokenizer-Eval Public
    OpenMOSS/MOSS-Audio-Tokenizer-Eval’s past year of commit activity
    Python 3 Apache-2.0 0 0 0 Updated Jun 16, 2026
  • MOSS-Audio-Tokenizer Public

    MOSS-Audio-Tokenizer is a Causal Transformer-based audio tokenizer built on the CAT architecture. Trained on 3M hours of diverse audio, it supports streaming and variable bitrates, delivering SOTA reconstruction and strong performance in generation and understanding—serving as a unified interface for next-generation native audio language models.

    OpenMOSS/MOSS-Audio-Tokenizer’s past year of commit activity
    Python 233 Apache-2.0 16 3 1 Updated Jun 16, 2026
  • Llamascopium Public

    Performant framework for training, analyzing and visualizing Sparse Autoencoders (SAEs) and their frontier variants.

    OpenMOSS/Llamascopium’s past year of commit activity
    Python 221 29 8 0 Updated Jun 16, 2026
  • OpenMOSS.github.io Public
    OpenMOSS/OpenMOSS.github.io’s past year of commit activity
    HTML 2 0 0 0 Updated Jun 8, 2026
View all repositories

Top languages

Loading…

Most used topics

Loading…