An AI that costs 20 times more than its competitor returned a white screen.
The cockroach drew a sunset in the ocean for 3 cents. The test was published by an evaluator who follows these models closely. In BridgeBench, the Claude Opus 5 lost to DeepSeek V4.
They built an automated business making $2.4M with Opus 5 at 18 - and came to Stanford to show how to repeat it
00:50 - Opus 5 automates an entire business in 60 minutes 21:34 - one agent replaces 50 Anthropic engineers 45:47 - one file in Claude Code launches a business making
When it lands, I'll be interested in adding split-screen multiplayer
I spent some of my usually-unused Opus 5 credits to do this in preparation.
DeepSeek-v4.1-Flash being nearly 90% cheaper than Opus 5 and GPT-5.6 Sol is wild.
If the UI/UX is still this close at that price, i can see why people are testing it. and the bigger question is how it holds up across more designs.
You have reached the end of the archive
All of Claude Opus 5