重度付费用户模型排名：Opus 5.5 居首，Gemini Argon 未测
一位每月花数千美元订阅 AI 的用户按从好到差给模型排序，称 Opus 5.5 全能、GPT Astra 做 UI 很差，并注明自己还没试过 Muse Wonder 和 Gemini Argon。
I'm spending thousands monthly on AI models and AI subs so this is how I would rate current models from order of best to worse (I haven't tried Muse Wonder nor Gemini Argon so I can't speak for those) - Opus 5.5: Goated, can do everything, great taste and aesthetic sense, ultimate UI builder - GPT Astra: Good for most stuff and hard non coding problems, absolutely ass at any UI development, horrible taste - Grok 4.7: Super fast, insane workhorse, beats K3 and GLM in all my benchmarks of onchain investigations both in speed and accuracy - Abliterated Large v2: IYKYK - Kim K3 and GLM 5.3 : Good workhorses and good to delegate non critical development and data analysis, a bit slow - GPT 6.1 SOL: LOL 🤣🤣 wtf is this trash, dont trust it for anything code related, good for data analysis or non development work