Gemini 4 Carbon 预览版首测：3D 生成与 UI 表现对比主流模型
第三方评测者对 Google DeepMind 的 Gemini 4 Carbon 预览版进行了初步测试，重点评估其在 3D 生成、创意提示及 UI 任务上的表现。结果显示该模型在部分场景优于竞品，但在细节处理上存在明显短板。
报道正文
Arena.ai (@arena) · 2026-10-28 19
First impressions of Gemini 4 Carbon’s preview by @GoogleDeepmind are in. @petergostev just put it through a long run of 3D generations, creative prompts, visual explainers and UI tasks, compared against Gemini 4 Argon, Claude Opus 5.5, GPT-6.1 Sol, Muse Wonderful, Kimi and DeepSeek.
Watch now to see an early preview of the latest model coming from @GoogleDeepMind.
0 Intro: Gemini 4 Carbon 0 First test (White House): Carbon vs Argon 1 Carbon vs Claude Opus 5.5 2 Carbon vs GPT-6.1 Sol 3 Carbon vs Muse Wonderful and Kimi 3 Why test with 3D generations 4 Playable 3D game: Carbon, Argon, Sol, Opus 6 Where Carbon falls short: gaps in detail 7 Stonehenge 8 Golden Gate Bridge 10 Golden Gate Bridge: GPT-6.1 Sol comparison 11 Carbon's creativity 15 Monet paintings transition 18 Van Gogh paintings world 21 Why creativity matters in a model 23 Another 3D prompt: Carbon vs Opus, Sol, Muse, DeepSeek 26 London through the ages 29 Istanbul 31 Harbin ice city 33 New test: visual explainers (how lightning works) 35 John Snow and the cholera map 38 UI work 39 Golden Gate Bridge completion animation 40 More UI generations 41 Ski resort site and ramen ordering 44 Verdict
