steve beyatte

Anthropic releases Claude 4- what do you think?

Anthropic just released two new models for Claude 4: Opus 4 and Sonnet 4. Let's find out how good they are.

👉 What are you using it for?

👉 What’s impressed you? What’s broken?

👉 How does it compare to GPT-4, Gemini, and others?

Drop your thoughts, screenshots, prompts, use cases below.

Let’s trade notes while it’s hot 👇

260 views

Add a comment

Replies

Best
steve beyatte

At first glance, Opus and Sonnet 4 do appear to be better than 3.7 at writing and coding. I just updated all of my projects to use Opus 4 and the responses are of a noticeably higher quality.

I have noticed that Gemini, GPT4*, and the others do not do well at writing feature tests so that's one of the first places I start. Sonnet 4 just did a really good job from the random old controller/service code I gave it.

Exciting times- good work Anthropic.

Jean Willame

Sadly Claude Opus 4 is on the MAX plan on Cursor :(

But I can already see more relevant content from my claude desktop project either.

steve beyatte

@jean_willame Dang it's still not out on Windsurf.

André J

@jean_willame  @steveb Cline gets everything day 1. No waiting

Parth Ahir

Claude 4’s Opus and Sonnet are really impressive, especially in code generation and reasoning tasks. The audio changelog feature is unique too. I find it slightly faster than GPT-4 on some coding problems but still catching up in broad creativity. Anyone else testing multi-turn conversations?

steve beyatte

@parth_ahir What's a multi-turn conversation?

Parth Ahir

@steveb A multi-turn conversation is when you and an AI go back and forth, with the AI remembering context—just like a real chat between people. (What we usually do with AI)

Ever since Claude 4 was released, I've been excitedly testing it all day. I found that using the same prompt, Claude 4 generates webpages that are more colorful, smoother in interaction, and richer in detail than before.

And web games built with Claude 4 also have impressive detail and smoother interactivity. For example, I used this simple prompt in @Trickle which now powered by Claude 4: "Build a fully functional Tetris game with clean UI, basic controls, score display, and level tracking." — and the result was a highly playable Tetris game, right out of the box.

And for the pricing, Sonnet 4 keeps similar pricing to earlier versions, while Opus 4 uses more tokens but offers significantly deeper detail and more comprehensive outputs at a higher price.

If you're curious how Claude 4 performs in real-world UI/web/game/app building, try vibe coding with Trickle and see the difference for yourself. 😉

Charlotte Taylor

what stood out to me most was their focus on marketing AI agents capable of :

  • working on long . sustained tasks ,

  • understanding nuanced , real-world contexts ,

  • and collaborating effectively with humans .

They've also stayed firm on safety and interpretability ,which shows they're serious about making AI a co-worker ,not just a flashy tool .

This feels bigger than just shipping features - it's about redefining the relationship between humans and machines . A new kind of production dynamic is taking shape , and it's setting the stage for a future where developers and AI can truly co-create.