Hook

Their other posts in the index, biggest breakout first.
All right boys, Opus 5 is officially out to every single paid member on Claude code. The craziest, most odd part though is that it has been ranking higher to their own benchmarks, higher than Fable 5 on every single benchmark. How does that make sense? To me, it's like Fable 5 was government mandated. The government banned it, right? And Mythos 5. Then they were doing like every week, find will extended, find will extended, and then they just finally released it to everybody. But somehow, Opus 5 is stronger than Fable 5, but it's just out. And these are their benchmarks, not mine. Like, I don't get it. Either way, they're saying this is their most behaved model. So on like misaligned behavior and what it's not allowed to do or it is allowed to do, it has basically behaved the best. So it's more Jarvis than Ultron. But the most important part here is the cybersecurity, right? It's saying that Opus 5 is barely, barely below Mythos 5 on vulnerability identification. But for exploitation, I would say like hacking, it's much, much dumber than Mythos 5. So like how is Mythos 5 still banned from everybody because of this like hacking part, not the finding part, but Fable 5 is supposed to be just as smart as Mythos 5, but that's allowed. And what they're saying now is that Opus 5 is as smart as Fable 5, basically. But oddly enough, and good enough for us, it's being priced the same as Opus 4.8, which doesn't make any sense. So Opus 5 is basically as smart as Fable 5 and in some instances, Mythos 5, but it's still not as smart as Mythos 5, but it's being priced the same as Opus 4.8, yet it doesn't have that government Ultron bureaucratic tape around it. It can just be released to everybody instantly. This is why I believe a lot of this stuff is just marketing. A lot of it is marketing. I'm not saying the government is marketing with them, but at the end of the day, it's just like, oh my god, they made Ultron and now I get to use Ultron. But then it's like next week, wait, this new model, Opus 5, is cheaper than Ultron, but it's just as smart as Ultron. Why wouldn't I just use that instead?