Four models, one neon pinball prompt: 6 minutes vs 2 hours
On the same neon pinball prompt, GPT-6 Luna finished in 6 minutes for about $0.08 and Sol in 17 minutes for about $1.20, while Grok 4.7 took over two hours and cost more.
6 MINUTES VS 2 HOURS FOR THE SAME PINBALL PROMPT IS CRAZY Four new models. Same prompt. XHigh. The task: build a playable neon pinball game Opus 5.5: 47:40 • ~$5.50 Grok 4.7: 2:05:49 • ~$13 GPT-6 Luna: 6:22 • ~$0.08 GPT-6 Sol: 17:26 • ~$1.20 Grok lost the plot and spent over two hours building pinball. Opus still mogged it on visuals - and our rough cost model puts Grok’s run at more than twice the cost. Which one would you actually play? Try the challenge via @aimlapi. Full scores at the end of the video.
This post quotes another post: x.com/noclipepe/status/2102572756353061250