Image Credit: MicrosoftMicrosoft launched its first in‑house text‑to‑image model, MAI‑Image‑1, in October 2025, debuting at #9 on the Arena leaderboard. In May, Microsoft released MAI‑Image‑2.5, its latest text‑to‑image model, which again secured the #3 position on Arena’s competitive leaderboard. Today, Microsoft introduced MAI-Image-2.6, its latest image generation model, which is now ranked #2 on the Arena leaderboard, only behind OpenAI's GPT-Image-2. MAI-Image-2.6 is now the #2 text-to-image model in the world - beating out Nano Banana, Meta, and Grok! Interested users can try out the MAI-Image-2.6 model for image generation on Arena.