Unverified50% confidenceReleaseExact time
Fudan University, together with the OpenMouse team, open-sourced the 5B image generation model Mouse Image, which natively supports ultra-long prompts of 4000 characters, can precisely render dense Chinese and English characters, and supports structured code control of image layout.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/29/2026
First Seen
Valid until: 11/12/2026
Sources
Gemini 4 Carbon Tops Global Rankings as AI Productivity and Open-Source Ecosystems Surge
bilibili炙心zx10/29/2026
Related Entities
Related Claims
UnverifiedImagen的核心架构基于大规模语言模型(如T5-XXL)作为文本编码器,配合级联式扩散模型进行图像生成74% similarUnverifiedMOSS-Image is a 5.1B text-to-image model released by OpenMOSS that supports rendering English and Chinese text in generated images and offers SFT and Base versions.74% similarUnverifiedCLIP由OpenAI与微软联合开发,通过对比学习在4亿张图文对上建立跨模态语义理解69% similarUnverifiedDALL·E 3 is the third-generation text-to-image model released by OpenAI, with significant improvements in text rendering and detail control.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1111421API
curl https://kongchang.com/api/v1/knowledge/claims/1111421MCP
get_claim(id=1111421)