Gemini 3.8 Flash vs GPT 5.6 Sol vs Claude Opus 5 vs Kimi K3.
Same brief. Four 3D rocket launches. Flash did not “fail at 3D.” It made a runnable scene. It just lost. Toy geometry. Hard white cone instead of smoke. One frame goes blank. $2.30 and 30 min in Goal Mode did not save
3D dot sphere made with Framer Agents and Opus 5 💫
Result from Opus 5 on the same prompt.
It was one shoted
Google and Mera Just released their best model.
Here are the results on real UI task. (Same prompt ) On ArtificialAnlys bench Muse is beating FABLE 5 - i want you to decide on real task. They both claimed to be either beating opus or at opus level. Result from Opus 5 is 👇
You have reached the end of the archive
All of Claude Opus 5