It must be some kind of psy-op by OpenAI to claim that Sol is anywhere near as good as Fable
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。
Claude 相关模型动态已经出现，适合跟踪能力变化、生态影响和后续可用性。
一位同时订阅了ChatGPT Pro和Claude Max的用户认为，声称OpenAI的Sol模型与Fable相当的说法是个“笑话”。他们指出，Claude的Opus 5与OpenAI的旗舰产品5.6 Sol价格相似，但输出质量大致相当。用户表示，尽管他们的立场在评论中被认为带有毒性，但他们需要花费相似的精力来纠正这两个模型，以获得可比较的结果。
I have a ChatGPT Pro subscription and a Claude Max subscription, and use both extensively for work. To claim that any model offered by OpenAI is even close in capability or problem solving ability to Fable is a joke to me.
To me, the most comparable Claude model to 5.6 Sol, OpenAI's flagship, is Opus 5. They have roughly equivalent price (ignoring the temporary promotions on Sol pricing), and in my experience, their output quality is about the same as well; I end up having to put in about the same amount of effort correcting them or giving feedback to achieve a product of comparable quality.
The main difference is in the kind of feedback I have to give; with Sol, I typically end up having to add details to its results, such as instructing it to address missing edge cases, or take a more thorough approach when it took a simpler shortcut to solve my problem instead. With Opus, it usually finds most edge cases for me without having to say anything; but it also goes beyond and keeps finding more and more things, of decreasing and often spurious relevance to my actual problem. My effort usually comes in the form of telling it to ignore those extraneous edge cases and focus on the core of the problem.
But when compared to Fable, neither can hold a candle. Among every task I've ever given any agent, Fable always takes the least amount of time, the fewest tokens, and needs by far the least number of warnings in the prompt or corrections to the output, compared to any other Anthropic or OpenAI model.
To me, to say GPT 5.6 Sol is anywhere close to Fable in any capacity, and not just a competitor to Opus with different tuning, is completely unfathomable to me. You pay twice the price for it and you get your money's worth. Sure it's expensive, and you can run through your weekly limits in hours, but you can't argue that it just works. I can't say the same about Opus or Sol.
Edit: wth is with all the toxicity in the comments jesus. i thought you guys would be appreciative of more anthropic support when everyone on this sub is always complaining about expensive limits and annoying opus.