4.6 still the GOAT
热度趋势
百分比基于当前可用热度信号，而非评论数或独立用户人数。
一位开发者在 Reddit 上表达了对 Opus 4.6 的强烈偏爱，称其为“GOAT”（史上最佳）。他们回忆说，4.6 通过理解复杂的请求并构建更好的解决方案，显著改善了他们对 AI 的使用。相比之下，Opus 5 被描述为需要大量的指导，并且故意误解指令，导致沮丧。该开发者在使用 5 遇到困难后，用 4.6 第一次尝试就成功完成了一项任务，并询问其他人是否仍在使用 4.6 或更喜欢 5。
I have very fond memories of Opus 4.6 being the greatest model at the perfect inflection point for my AI usage. It felt like overnight I was able to go from reviewing every line of code and worrying that I was going to be grossly misunderstood and destroy a feature to barely having to articulate what I needed to an insanely intelligent partner that would then build something better than I had originally envisioned.
Fast-forward to Opus 5 and I feel like I’m back to having to spoon feed every little step to something that intentionally tries to misunderstand me then gaslights me telling me it did what I asked. I spent hours today trying to build what should be a fairly simple listserv aggregator using Opus 5 and it just wasn’t working. I was getting very frustrated so I decided to switch back to 4.6 and I swear to God it got it right the first try and then it told me why 5 wasn’t able to do it.
I guess I’m just addicted to frontier models and feel like I need to be using the newest thing, but I see no reason to use anything other than 4.6 for routine projects until a newer model proves me wrong. Fable is amazing but difficult to justify for daily use at the cost.
Anyone else still daily driving 4.6? Anyone actually prefer 5? If so, why?