museIMAGE
Users test Meta Muse Wonder's image generation with complex prompts
A user tested Muse Wonder with complex prompts like an emperor riding a polar bear and pelican, and a Sun Wukong flying a plane, plus a candy question; the candy question was answered correctly and the compositions were mostly reasonable, but details were rough, with the model rated near Gemini 3.8 Flash but slightly smarter.
Leo · Building w/ AI
@leo114119Archived 22 days agoOriginally posted 2026.10.13

Leo · Building w/ AI @leo114119Originally posted 2026.10.13
🎬 网友实测分享:meta muse woder 综合大杂烩测试 meta muse woder 大模型生成的始皇坐白熊再骑鹈鹕,再坐孙悟空开飞机加糖果题 糖果题对了,画面也基本上都合理,但细节处理很粗 感觉就是一个gemini 3.8 flash的水平 但会更聪明一点
x.com/leo114119/status/2109794033723949078Archived 2026.10.14 10:30