Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
MiniGPT-4minigpt-4.github.ioMiniGPT-4 is a vision-language model that aligns a frozen visual encoder with the Vicuna large language model using a single projection layer. It enables capabilities like detailed image description, website creation from handwritten drafts, and story generation inspired by images.
Parse Score
