Key takeaways:
- New AI model selection button is live
- Claude 3.7 Sonnet reasoning model is now available
- UI is more streamlined and easier to use
So, Perplexity just dropped a couple of new updates, and honestly, they flew under the radar. No big headlines, no official YouTube announcement—just a quiet mention on Discord. But these tweaks? They’re actually pretty sweet and make using Perplexity way smoother.
The main thing: you can now pick your AI model right from the prompt screen, no more digging through settings. Plus, Claude 3.7 Sonnet reasoning is here, and it’s got a whole new way of thinking. I’ll break down what’s changed, how it works, and why you might care.
Table of Contents
- The new model selection button makes life easier
- Claude 3.7 Sonnet reasoning model brings a new way to answer
- UI changes make everything less confusing
- How the Claude 3.7 Sonnet reasoning model actually works
- Why these updates matter for everyday users
- Table: Quick comparison of Perplexity AI modes
- What’s next for Perplexity?
- FAQs
- How do I pick the new AI model in Perplexity?
- What’s special about Claude 3.7 Sonnet reasoning?
- Can I use the new model in follow-up questions?
- Is Deep Research different from Reasoning?
- Where can I find more tips for using Perplexity?
Before this update, picking which AI model to use in Perplexity was a pain. You had to go into settings, hunt for the model selector, and then get kicked back to the homepage. Now, there’s a button right on the main prompt screen. It’s labeled “Auto” by default, but you can click it and pick from the available models: Sonar, GPT-4.5, GPT-4o, Claude 3.7 Sonnet, Gemini, and more coming soon.
You can also choose between different response modes:
- Auto: Perplexity decides which model and method to use
- Reasoning: The AI “thinks” more like a human, breaking down questions step by step
- Deep Research: Exclusive to Perplexity, this mode pulls from tons of sources for detailed answers
If you’re curious about how to get the most out of your device or want to know how to tweak other settings, check out how to speed up Windows easily.
Claude 3.7 Sonnet reasoning model brings a new way to answer
The big headline for AI nerds: Claude 3.7 Sonnet reasoning is now an option. This isn’t just a regular model—it’s a “thinking” model. When you ask it a question, it doesn’t just spit out an answer. It breaks down the question, analyzes each part, and gives you a step-by-step response.
For example, if you ask, “How many R’s are in Strawberry?” it’ll actually count each letter and explain the process. Or if you throw a joke question at it like, “Why don’t we visit the Sun at night when it’s not hot?” it’ll give a serious, logical answer, explaining why the Sun never “turns off.”
UI changes make everything less confusing
The interface got a subtle but important tweak. Now, the model selection button is always visible when you start a new thread or even in follow-up questions. No more hunting through menus or getting lost in settings.
This update makes it way easier to decide how you want your question answered. Want a quick response? Stick with Auto. Need something more detailed? Try Reasoning or Deep Research.
If you’re all about customizing your workflow, you might also be interested in how to change Chrome’s default download location.
How the Claude 3.7 Sonnet reasoning model actually works
Let’s talk about what makes this “reasoning” model different. It’s not just about finding info—it’s about showing its work. When you ask a question, you’ll see the AI go through each step, almost like watching someone solve a puzzle.
- For logic or math questions, it’ll break down the problem piece by piece.
- For silly or trick questions, it’ll still give you a thoughtful answer, even if the question itself is a joke.
- It’s less likely to “hallucinate” or make up answers, since it’s pulling from more sources and showing its logic.
If you’re into troubleshooting or fixing things on your own, you might like how to fix USB ports on Android phones.
Why these updates matter for everyday users
If you use Perplexity for research, homework, or just random questions, these updates save you time. You don’t have to jump through hoops to pick the right model. The reasoning model is great if you want to see how the AI thinks, not just what answer it gives.
And if you’re working with big files or need to transfer stuff, check out how to transfer big files to USB stick.
Table: Quick comparison of Perplexity AI modes
| Mode | How it works | Best for |
|---|---|---|
| Auto | Picks model and method automatically | Fast, general questions |
| Reasoning | Step-by-step logic, human-like | Complex or tricky questions |
| Deep Research | Pulls from many sources, detailed | In-depth research, accuracy |
What’s next for Perplexity?
The devs are hinting at more models coming soon. The UI is getting cleaner, and the focus is on making the platform more user-friendly. If you want to keep your browser running smooth while you’re using Perplexity, see how to make Chrome consume less memory.
FAQs
How do I pick the new AI model in Perplexity?
Just click the model selection button on the prompt screen. No need to dig through settings.
What’s special about Claude 3.7 Sonnet reasoning?
It actually “thinks” through your question, showing each step, not just giving a quick answer.
Can I use the new model in follow-up questions?
Yep, the model picker is available in follow-ups too.
Is Deep Research different from Reasoning?
Yes, Deep Research pulls from tons of sources for a more detailed answer, while Reasoning focuses on logical steps.
Where can I find more tips for using Perplexity?
Check out guides on speeding up Windows or fixing USB issues for more tech help.
🧑💻 That’s the rundown—new model picker, smarter answers, and a smoother experience. Give it a spin and see how it changes your workflow!
