Same-prompt Blender modeling test: author says GPT-6.2 Sol 'was not close'

iamfakhrealam · x · 2026-11-22

From iamfakhrealam's roundup: @vipinxtwt ran the same prompt to build a model in Blender from scratch and claims the competing GPT-6.2 Sol result wasn't close — an informal comparison, no benchmark data attached.

Original post →

More from Models

Models channel →