会写字、懂排版、能分镜：HiDream-O1-Image-1.5 全能图像生成能力解析
Quick Answer
HiDream.ai's HiDream-O1-Image-1.5 model achieves SOTA performance, ranking first in China's Text to Image Leaderboard, surpassing major competitors like OpenAI and Google.
Quick Take
This model showcases enhanced capabilities in image quality, semantic adherence, and complex scene generation, marking a significant advancement in commercial applications across various industries.
Key Points
- HiDream-O1-Image-1.5 ranks first in China's Text to Image Leaderboard with 1265 ELO.
- The model excels in generating high-quality images with complex layouts and multi-subject consistency.
- It supports various commercial applications including advertising, brand design, and game content.
- The Unified Transformer architecture enhances its performance in multi-modal tasks.
- HiDream-O1-Image-1.5 demonstrates strong capabilities in text rendering and narrative generation.
Source Excerpt
近日，智象未来（HiDream. ai） 全新推出的商用版图像生成模型 HiDream-O1-Image-1. 5 再次实现 SOTA，在全球知名独立 AI 模型评测与分析平台 Artificial Analysis的文生图榜单（Text to Image Leaderboard）上，一举登上中国图像生成模型第一，成为评分仅次于OpenAI的中国大模型公司，超越Google Nano Banana 2（Gemini 3. 1 Flash Image Preview）、NVIDIA Cosmos3-Super-Text2Image 和字节跳动的Seedream 4. 0等国内外大厂的主流图像生成模型。 半个月前，智象未来 HiDream-O1 系列的开源模型 HiDream-O1-Image-Dev-2604 刚刚登顶文生图榜单开源模型全球第一。 数周之后，HiDream-O1-Image-1. 5再次进入全球文生图大模型榜单前三，连续登顶不仅印证了智象未来在图像生成大模型上的硬核实力，更标志着其已在视觉生成大模型的角逐中稳居全球第一梯队阵营。
Artificial Analysis 的 Text to Image Leaderboard 采用匿名对比、用户投票和ELO动态排名机制，尽量减少品牌认知对评测结果的影响，更接近真实用户在开放生成场景中的偏好判断，在这一专业评测体系下，HiDream-O1-Image-1. 5在超过4000 个样本对比中取得1265 ELO。 …
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from 雷峰网 AI
See more →
刚刚，GPT 5.6 发布会上，OpenAI 暴露了哪些 Agent 技术路线？
OpenAI's GPT 5.6 integrates ChatGPT and Codex, introducing a for complex task execution, with models Soul, Terra, and Luna for efficient workflow management. The release emphasizes task orchestration, contextual understanding, and robust security measures for enterprise applications.
