Opus 5 vs DeepSeek v4.1 Flash tested both models with same frontend prompt at highest reasoning available but results came out really different > opus took 80 minutes to complete and costed $20 > v4.1 flash took 60 minutes and costed $0.2 which one did better here?

deepseek overthinking hasn't been fixed at all model is fast but it keep over thinking and ends up taking long time opus 5 took 80 minutes and i wasn't able to scroll site at all i could have asked it to fix it but it was one shot test for both models here's prompt: create a website for frontend devloper morden and stunning creative should have 3d models ,parallax scrolling and lenis scrolling effects and much more and smoother user interface experience use advance tech stack and write code i dont know what you do but i want that test yourself and share outputs if you can

Claude cooked by chinese model 😂😂😂😂

yeah it's weird how Claude models are super slow

Flash modle mogged Opus 💀

imagine 20 vs 0.2 no comparison bro .. v4.1 flash is just different .. too good.

Haha yess

the quality and cost efficiency of v4.1 flash is so impressive legend

nothing can beat cheap and better model

👀 DeepSeek mog?

I like what opus made but it didn't made anything to let me scroll in 80 minutes

I liked the one on the left; the fonts look more interesting

In the span of a week, OpenAI and DeepSeek are driving down Anthropic's IPO valuation by 30% to 50%.

😭😭😭

Opus 5🫠🫠🫠

what is wrong with opusies

didn't expected new DeepSeek model today

it was out few days ago already

More complex tests are needed to understand the capabilities of DeepSeek v4.1 Flash

running those 3d one for sure so let's see what it could do

$20’ for this???

yeah

If the final output was equally good, V4.1 Flash wins easily..... 100× cheaper and faster

I'd choose v4.1 anytime, it's efficient and cost effective

Chinese models are going crazy rn

DeepSeek is looking really good for its pricing. Ig I should give it a try 😸

Flash wins

that $20 vs $0.20 is the whole story. even if quality was tied, finishing 20 mins faster for 100x cheaper is a brutal flex for flash

Deepseek

what about astra?

Can I use it on work projects though? Or will they sell all my sensitive data?

Flash model that costs absolutely nothing destroyed opus that costs 2 kindneys to run and takes 15 years to do it

Wow 🔥🔥🔥

the metric is cost to accepted patch, not cost to first render. set the same scroll, responsiveness, and style checks, then let each model iterate until it passes. production routing starts there.

Opus 5🫡

Deepseeks first multimodal model really showing of wow!

Ran a similar Opus vs smaller-model bakeoff on a frontend ticket. Fast-and-cheap still blew the design system; slow-and-pricey respected the tokens. I now score on "matches existing CSS vars" first, not wall-clock.