+1 (415) 347-6981Get started
Product & Engineering
Brand & Design
AI & Automation
AI Security
AI Labs
Ongoing
Fixed prices · You own the code · Proposal in 24h hello@startrise.io →
AI Model Release Tracker China · updated
  • 1Releases tracked
  • Oct 16Latest · MOSS-Image (Base and SFT)
  • 100%Open-weight
  • CNChina

OpenMOSS is a China-based open-source AI research group that builds open models for language, speech, video and embodied intelligence. It operates under the Shanghai Innovation Institute (SII), works closely with Fudan University and MOSI Intelligence, and is led by Prof. Xipeng Qiu. Its open projects include the MOSS bilingual dialogue model, MOSS-TTSD for multi-speaker dialogue synthesis and AnyGPT. The tracked release, MOSS-Image (Base and SFT, October 2026), is a 5B text-to-image model with JSON layout control, strong text rendering and prompts up to 4K tokens, released with Apache-2.0 weights.

OpenMOSS

China openmoss.ai

1 release latest Oct 16, 2026

  1. MOSS-Image (Base and SFT)

    5B text-to-image DiT with JSON layout control, strong text rendering and prompts up to 4K tokens; Apache-2.0 weights.

    • Image
    • Open weights
    Source

OpenMOSS FAQ

What is OpenMOSS's latest AI model?
OpenMOSS's most recent release in the Startrise tracker is MOSS-Image (Base and SFT), released Oct 16, 2026.
How many AI models has OpenMOSS released?
Startrise tracks 1 notable OpenMOSS release since January 2025, from MOSS-Image (Base and SFT) (Oct 16, 2026) to MOSS-Image (Base and SFT).
Does OpenMOSS release open-weight models?
Yes. 1 of 1 tracked OpenMOSS releases (100%) ship with downloadable weights; the most recent is MOSS-Image (Base and SFT).
Where is OpenMOSS based?
OpenMOSS is based in China.