Prototyping iOS apps with Claude Opus 5

John Rodrigues demonstrates prototyping an iOS card interaction with Claude Opus 5, Anthropic's new model that costs roughly half as much as Fable while offering comparable capability. Its 750K-word context window lets it retain information across long interactions, though the final 20% of design polish still needs human judgment.
Chapters
So welcome to another video. In this video, we're going to talk about prototyping iOS apps with the Opus, Claude's Opus 5. So Anthropic recently launched their new model called Opus 5, which is equally good and capable as Fable, but half the price. And I've been— I shared it here as well in some of the shorts, and I've been researching about the other models like KIMI K3, which is like so cheaper and equally getting good as some of the frontier best models, right? With a gap of like 3% and stuff like that. And I believe this is Anthropic's answer to it. So the price is definitely lower now, but it's equally capable as Fable. So Fable is still Anthropic's best model. And so, but I think for most people, you don't need that level of capability, right? So what I did was, so at the Humani AI Studio, we do a lot of like prototyping and design system work and things like that. So I thought to experiment and apply to one of the— so I thought to experiment and apply to creating this card interaction here. So to be honest, for the first shot, it kind of did a decent job. I was like surprised, but definitely not that last 20%, right? So I think that last 20% of product design, design engineering work still kind of remains with that human. And then that human judgment and taste still plays an important role. But it's It's getting pretty good day by day, but I was able to create this interaction with like using iOS design systems, animations and things like that, and using the existing components and stuff. It pulls pretty well. Apple's design system is like really good. But so yeah, and really good. So I was able to do that. So a few things to know about the Clouds Open 5 is it has, it was launched on July 24th. Very, uh, 2026. And, um, uh, it holds, it holds a longer context windows and then 7,000, uh, 750K words and 1 million token windows. The thing is like the model, like almost all, most of the models, I'll be able to tell the same that they'll hold longer context now, remember stuff. But Opus 5 does really good job with the remembering. And you probably have this issues before where the AI, the models, the chat windows, right? You talk and it forgets the context, but now the Opus 5 can hold a longer context. And most of the models are definitely kind of, you know, giving this kind of a performance and stuff like that. But I think the biggest takeaway from this release is it's equally capable as Fable, and Fable is one of the best models, yet half the price of the Fable. And that's like the biggest takeaway from this recent update. So at Human AI Studio, I've been developing an agentic design system, and you probably have been following some of the series. If not, check those agentic design system. And then I've developed a lot of like components and like run loops and stuff. And loops is an interesting concept that recently has become part of our workflows and it needs a longer context and remembering and things like that. So it's been super helpful and I've been applying for that here. So make sure you kind of like, I'll publish and I'll publish more videos. I'm gonna publish like more videos and stuff like that soon. So make sure you are subscribed to this YouTube channel and if you are watching this and subscribe to the publication. But one thing I can see is like, you know, if we even saw the Thinking Machine published, Thinking Machine published Inkling model, which makes like fine-tuning of this model easier. KIMI K3 was like so cheaper with so much possibility, so much capabilities. And now Opus 5, which like half the price of Fable. It's definitely getting clear that to me that the, you know, the price should not be a problem to run this model and companies spend like millions of dollars, you know, the pricing of the tokens and stuff like that is definitely coming down. I think that would definitely increase the adaptation. But I think the, in terms of like AI slop, right? I think that's also, I wouldn't say completely, but it's getting reduced. Like the output is not that bad like used to be before and it's going to get better with time. But the thing about, the thing about, I still bet on like, but the context is still the king, right? The models can be the most powerful model, but the context without that, it's not good. So like the second brain, you know, is something is like very important and also memory layers and then design systems for product and design engineers and stuff. Those things are still very important. And I believe like design system's gonna be the core, you know, infrastructure of product teams will operate. That's what I'm like developing. HumanAI Studio. The companies will have its own context and stuff. So, so, you know, you can't, so you can't bake in a context in a model to that company. So I think that's going to be the next big opportunity in the space of AI. But I think Opus 5 is pretty good. And I would say that the, the price is coming like down with all these AI models and stuff, and it's getting more accessible and stuff. So yeah, that's about it for this week video. And then if you enjoyed this video, make sure you subscribe to this YouTube channel, check out the resources in the description, and all the good stuff. I'll see you in another video.
More from John

Building Orb AI in Public: Episode 6 Accelerator & Creator Focus #creator #orbiai

Jev: The Decision Layer Boosting AI Retrieval from 72% to 95%. #Jev #AIDecisionLayer #GenerativeUI

Design Engineer's Ultimate Tool Library: designengineer.tools #designengineer #designsystems







