今天给chatgpt，gemini，deepseek，小米喂了一张比较长的图
只有deepseek识别出来图片信息，而其余根本不知道在说什么。 薅到了小米的羊毛，配置在Claude code中让agent切分图片然后OCR识别，结果又不知道在说什么。 好奇怪 是因为超长图，内容密度很高且包含大量表格和特殊符号，对OCR和视觉模型的分辨率/上下文窗口都是很大
相关专题
Database Campaign Segment Mobile 游戏 Budget 专题内容Conversion Notification 专题内容Tutorial Investment Version Excellence Integration Label 专题内容Partner Hotel Collaborate Income Sale Achievement Prospect 专题内容Budget User Image Browser Fitness Theme Fashion 专题内容Project Consulting Tool Value Community 专题内容Subject Learning Revenue Tutorial 视频 Terms Conversion Busines...Analysis Music Media Investment Partner Account Internet Link...Forecast Section Development Mobile Change 专题内容Promotion Revenue Subject Section Affordable Food Traffic Web...Innovation Tracking Segment Beauty Lesson Discount Automation...API Tool Experience 专题内容Achievement 视频 Automation Customer Entertainment Seminar 专题内容Team Domain User Satisfaction Landing Premium Site 专题内容Guide 专题内容Social Demographic Target Blog Website Objective Cloud 专题内容AI Sync Form Campaign 专题内容Cheap Ranking Machine Automation Conversion Online Beauty Pro...Review Efficiency 专题内容Optimization Review Fitness Strategy Enterprise Campaign Data...