Microsoft’s MAI-Image-2.5 has achieved notable rankings in the Artificial Analysis Image Arena, securing #2 in Text to Image generation and #3 in Image Editing, following only OpenAI’s image models. This model, capable of generating images with a maximum output resolution of approximately 1MP and a flexible aspect ratio, allows users to create detailed visuals and modify existing images based on complex editing instructions. Both MAI-Image-2.5 and its faster, lower-cost variant, MAI-Image-2.5-Flash, are available via the Foundry API, MAI Playground, and are integrated into Microsoft applications like PowerPoint and OneDrive. These capabilities highlight Microsoft AI’s commitment to advancing image generation and editing technologies.

Google: Google is a technology company that develops AI systems including the Gemini family of models. Its Nano Banana 2 image model ranks among the top performers in recent text-to-image and image-editing evaluations.
OpenAI: OpenAI is an AI research organization focused on developing advanced generative models. Its GPT Image 2 and related models currently lead multiple categories in independent image benchmarks for text-to-image generation and editing.
Microsoft: Microsoft is a major technology company that develops software, cloud services, and AI technologies through its Microsoft AI division. The company recently introduced MAI-Image-2.5 as part of its growing family of image generation and editing models.
GPT Image 2: GPT Image 2 is OpenAI’s advanced image generation and editing model. It currently holds the top position in text-to-image rankings and second place in image editing on the Artificial Analysis Image Arena.
MAI-Image-2.5: MAI-Image-2.5 is Microsoft’s latest image model capable of both text-to-image generation and image editing at up to roughly 1MP resolution with flexible aspect ratios. It achieved #2 ranking in text-to-image and #3 in image editing on the Artificial Analysis Image Arena leaderboard.
Nano Banana 2: Nano Banana 2 is Google’s image model, also referred to as Gemini 3.1 Flash Image Preview. It ranks fourth in both text-to-image generation and image editing categories behind the leading models from OpenAI and Microsoft.
MAI-Image-2.5-Flash: MAI-Image-2.5-Flash is the faster, lower-cost variant of Microsoft’s MAI-Image-2.5 model designed for efficiency. It placed #8 in text-to-image and #6 in image editing on the same independent leaderboard.

{“Availability”: “MAI-Image-2.5 and its Flash variant are accessible via the Foundry API, MAI Playground, and integrated into Microsoft applications including PowerPoint and OneDrive.”, “Benchmarking”: “Models are evaluated through blind user votes in the Artificial Analysis Image Arena using an Elo rating system for both text-to-image generation and image-editing tasks.”, “Model Capabilities”: “MAI-Image-2.5 supports up to 32K token context and handles complex editing instructions alongside original image generation.”}