A user has developed a blind arena page where six different AI models are given the same prompt to build a voxel space station. The models' creations are then voted on before their identities are revealed. This test included models such as Opus 5.5, Sonnet 5.5, Fable 5.1, Muse Wonder 1.0, GPT-6.1-Sol, and GPT-6-Astra, with build times ranging from under five minutes to over two hours. Claude was instrumental in building the arena page and ensuring the creations were compatible across devices. AI
IMPACT Highlights performance differences between AI models in creative generation tasks.
RANK_REASON User-created comparison tool for AI models.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →