MIRAGE

Compare Text-To-Image Models Side by Side

Multi-model Interface for Reviewing and Auditing Generative Text-to-Image AI

Compare Models

Generate images from multiple AI models at once and compare their outputs side by side.

Detect Patterns

Identify patterns, biases, and inconsistencies across different text-to-image models.

Contribute

Share your findings and help improve AI systems by providing valuable feedback.

Select a Language to Start

Research Papers

Consider citing both papers if you use MIRAGE in your research

Seeing Twice: How Side-by-Side T2I Comparison Changes Auditing Strategies

Matheus Kunzler Maldaner, Wesley Hanwen Deng, Jason I. Hong, Ken Holstein, Motahhare Eslami

ACM Collective Intelligence 2025

Demonstrates how side-by-side comparison interfaces change user auditing strategies, revealing model personalities and bias patterns in text-to-image AI systems.

Read Paper

MIRAGE: Multi-model Interface for Reviewing and Auditing Generative T2I AI

Matheus Kunzler Maldaner, Wesley Hanwen Deng, Jason I. Hong, Ken Holstein, Motahhare Eslami

AAAI Conference on Human Computation and Crowdsourcing 2024

Introduces the MIRAGE system architecture and methodology for comparative analysis of multiple text-to-image models in a unified interface.

Read Paper