POPA ReconTop 10 Text-to-Image Models
Top 10 Text-to-Image Models Microsoft's MAI-Image-2.6 landed at #2 on Arena's Text-to-Image leaderboard with 1,336 points, 45 behind GPT Image 2 (Medium) at 1,381 and 20 ahead of Grok Imagine at #3, per Arena's head-to-head user votes. The jump is the story. MAI-Image-2.5 sits at #10 with 1,256 - one release moved Microsoft 80 points and eight places, past Meta, Google and ByteDance. For a company that spent three years reselling OpenAI's image models inside Copilot and Bing, having its own model within 45 points of its partner's is the point: the very labs that fund OpenAI are now shipping…
The excerpt is quoted from a public Telegram channel, which owns its text.